Report - 2025 01 22 A32494 D275397 Ago Inaugural Report Of The Washington State Artificial Inte

Issuer: Congressional materials
Document type: Report
Date: 2025-01-22
Case: 2025 01 22 A32494 D275397 Ago Inaugural Report Of The Washington State Artificial Inte

Summary

The Inaugural Report of the Washington State Artificial Intelligence Task Force, issued by the Washington State Attorney General's Office and dated December 2024. It opens with a letter from Attorney General Bob Ferguson and lists the Task Force's 19 members, established under Substitute Senate Bill 5838. The executive summary states that the Task Force met in July and October and held a voting meeting on December 16 to approve its first recommendation. That recommendation asks the Legislature to amend Washington's child sexual abuse material (CSAM) law by removing the requirement that fabricated images depict a minor who is identifiable. A background section describes generative and predictive AI, policy concerns such as algorithmic bias, deepfakes and energy demand, and state and federal actions including Executive Order 24-01. The report closes with endnotes citing its sources.

Summary drafted by a model from the document's text below and checked by script against that text before publication. It is a navigation aid, not a reading of what the document proves. Where AI is used

Cite this document

“Report - 2025 01 22 A32494 D275397 Ago Inaugural Report Of The Washington State Artificial Inte.” Original at app.leg.wa.gov, dated 2025-01-22. Archived at Pandemic Darlings, https://pandemicdarlings.com/source-documents/2025-01-22-2025-01-22-a32494-d275397-ago-inaugural-report-of-the-washington-state-artificial-inte/, SHA-256 2d8b7dec6d5f9f9826bb32432344cd801d48f14d248af09d7ff85d14464f5ecc.

Full text

Inaugural Report of the
Washington State Artificial Intelligence Task Force

Washington State Attorney General’s Office
December 2024

1
Letter from Attorney General Bob Ferguson
Dear Washingtonians,

Washington has a long and deep history as a leader in technological
innovation. Earlier this year, I asked the Legislature to establish
an Artificial Intelligence (AI) Task Force, administered by the
Attorney General’s Office. This Task Force must “…assess current
uses and trends and make recommendations to the Legislature
regarding guidelines and potential legislation for the use of artificial
intelligence systems…” The scope of work for the Task Force is expansive and important as
this technology rapidly develops. AI can provide benefits to people and communities across
the state, and it can also cause harm if developed, deployed, or used without appropriate
protections.

The Task Force brings together leaders and experts from the public and private sectors,
legislators, community members, and others to share their experience and knowledge. The
Task Force convenes eight subcommittees to discuss the application of AI in specific sectors
and identify guidelines and recommendations that appropriately balance protections for
Washingtonians with continued innovation and development. Anyone with experience or
expertise to share is welcome to join the subcommittee discussions.

This report is the first interim report from the Task Force, with information from the initial
Task Force meetings and subcommittee meetings. Over the next two years, the Task Force will
issue two additional reports with more information and recommendations for policy makers to
consider. I hope you find this information helpful.

I look forward to the Task Force’s continued work to ensure Washingtonians benefit from the
potential of AI and from important protections.

Sincerely,

Bob Ferguson
Washington State Attorney General

2
oN,
Contents °—

Letter from Attorney General Bob Ferguson.........csessssssscsssesssseseseesees 2
Task Force M@MDEeIG........cscsssscsssssessssssssscssssssssssseesssssesecsssesecsssssessesssesecsesees 4
EXO CUTIVE SUIMIMALY.......ccscssssssssssssscsssecscscscscsescscsesescsescsesssessessseeseesseesseesseeses 5
BACKQFOUNC ........csecsssssssssesssscscsesesssssscscsesssssscsescscsesessesesessseseesesesesescsesseseseseaeeees 6
Task Force WOrK .......cscsssessssssssscsssssscsssssecsssssesscsssescscssssscsssecscscsesscsesesessensees 10
Task Force ACtiVItieS .......cceesesssesessssesecsssssessssssesscsssssesessssesesssseseesesssesseacees 12
ENGQAGEME NL ........scescscsssssssescscsssssscssscsssssscecsssssscscsseessescessasseseeecasaccensasscsesees 14
RECOMMENAATION .........ccssessssesecscssssssssssssscsssssssscsssesscsssesssscassesecssassesesseseeees 15
LOOKING ANOad.........ceesessssssssssessssssescsessssssescsescsssssseeseseseseeseeseseacseeseesesesees 16
APPeNAIX 1: 5838 FULL TOXtececccccssssssssssssessssssesssssessssssesssssssssssssssssseesessseeesssseessseeeesssees 18
ADPD@NAIX 2: TASK FOPCE BYIQWS vessecsssssesssssssssssssssssesssssstsssssssessesssusessssseessesssuesssssneeees 22
Appendlx 3: TaSK FOrCe VOTING RECOM eseecssssssssssssesssssesssssssssssssssssssssssssessseessssees 28
Appendix 4: Al Policies In Other JULISCICTIONS ..ssecsssssessssssssssssesscssessssssesesssessesssees 29
ADPeNAIX 5: ELECTION GUICANCE veesessssssssssssssesssssessssssesssssssssssssssssssssessecessssesesssecessssees 43
EMOTES ceessssssssssssesssssssescssssssssssssessssssssesssssessssssusessssuscsesssssssssssuseesssusesessssseessssusesessseeeease 45
On

Task Force Members
The Task Force consists of 19 members representing agencies and entities specified by the Legislature.

• Sen. Joe Nguyen, D-White Center
• Sen. Matt Boehnke, R-Kennewick
• Rep. Clyde Shavers, D-Clinton
• Rep. Travis Couture, R-Allyn
• Sheri Sawyer — Deputy Director of Policy & Outreach, Governor’s Office
• Yuki Ishizuka — Senior Policy Analyst, Attorney General’s Office
• Scott Frank — Director of Performance and IT Audit, State Auditor’s Office
• Katy Ruckle — State Chief Privacy Officer, Washington Technology Solutions
• Dr. Magdalena Balazinska — Director, Paul G. Allen School of Computer Science and Engineering,
University of Washington
• Kelly Fukai — Chief Operating Officer, Washington Technology Industry Association
• Ryan Harkins — Senior Director of Public Policy, Microsoft
• Dr. Tee Sannon — Technology Policy Program Director, ACLU-Washington
• Vicky Tamaru — buildJUSTLY
• Paula Sardinas — WA Build Back Black Alliance
• Leah Koshiyama — Senior Director of Responsible AI & Tech, Salesforce
• Crystal Leatherman — Washington Retail Association
• Vacant – Hospitality industry representative (formerly Montana Miranda, Washington Hospitality
Association)
• Cherika Carter — Secretary Treasurer, Washington State Labor Council, AFL-CIO
• Chief Darrell Lowe — Redmond Police Department

4
Executive Summary
In response to rapid developments in artificial intelligence (AI) technology, Attorney General Ferguson
requested, and the Legislature passed, Substitute Senate Bill 5838, establishing the Washington State Artificial
Intelligence Task Force. The Task Force brings together leaders from the technology industry, private sector,
government agencies, and community to develop principles for AI development, deployment, and use, and
recommendations for policymakers to protect Washingtonians from the potential harms of AI, including bias
and discrimination, while supporting innovation and maximizing the benefits of AI to Washingtonians.
The Task Force met for the first time in July, and again in October. These meetings were primarily focused
on establishing the Task Force’s organizational and decision-making structure, defining key focus areas,
reaching out to relevant stakeholders and communities, and beginning to develop policy recommendations.
On December 16, the Task Force held a voting meeting to approve the first recommendation. The Task Force
will continue to meet quarterly, and its eight subcommittees will convene regularly for focused substantive
discussions to develop recommendations, initially focusing on developing AI principles and recommendations
related to transparency.
2024 Recommendation
This initial report includes one recommendation, asking the Washington State Legislature to amend a statute
enacted last year to strengthen Washington’s law on child sexual abuse material (CSAM) by removing the
qualification that fabricated images, including those created with the use of AI, of a minor engaged in sexually
explicit conduct include the depiction of a “minor who is identifiable.”

5
Background
AI is, in the broadest sense, the ability of technology to use information to learn and improve outputs to do
tasks generally associated with human intelligence or capacity. In the past several years, advancements in AI
technology received significant attention from policymakers and the public seeking to understand and respond
to the technology and its impacts. There are many different legal definitions of AI that serve different purposes,
including governing specific uses of AI or in specific sectors.
While AI has received a lot of attention recently, AI technology has existed for decades, and familiar functions
such as grammar check, voice to text technology, voice and facial recognition technology all use AI. However,
OpenAI’s release of ChatGPT in November 2022 accelerated advancements in AI technology. Within two
months of its release, ChatGPT became the fastest growing consumer application in history, with an estimated
100 million active users by January 2023.1
ChatGPT is an example of generative AI, or GenAI, technology that uses large amounts of data, called
“training data” to recognize patterns and produce outputs in response to user queries, and can generate text,
images, and videos. GenAI can provide benefits like summarizing meeting notes, setting reminders about
deliverables, streamlining tasks, research assistance, contract management, and enhancing user experiences
and services with tools like chatbots.2
Predictive AI, a different type of AI technology with a broad range of uses, also learns from data to predict
outcomes and behaviors. Predictive AI can improve the performance of large-scale operations, including
marketing, manufacturing, risk management, fraud monitoring, and many other functions.3
AI Uses
Algorithms powered by AI enable personalization of the
content and ads we all see on social media platforms and
the websites we visit,4 personalized options for viewers on
streaming services,5 increased efficiency of home heating
systems or other “smart” devices by identifying patterns in
use and time of day,6 and determines the quickest route to a
destination.7 In home AI-driven virtual assistants, such as
Apple’s Siri, Amazon’s Alexa, and Google Assistant, utilize
natural language processing (NLP), a form of AI, to understand
and respond to user queries, helping users manage tasks such
as setting reminders, playing music, and controlling smart
home devices.8
AI algorithms can be used to automate processes such as
decision-making, in which an algorithmic model makes
decisions based on large amounts of data. The algorithms need
to be “trained,” which means they must process large volumes
of data to predict outcomes and behaviors in which their decision-making is based.9 This technology has a
broad range of applications to automate tasks such as fraud monitoring for financial institutions by analyzing
transaction patterns to identify suspicious activities in real time.10
The business sector deploys AI for data analysis and decision-making.11 Machine learning algorithms process
vast amounts of data, identifying patterns and generating insights that inform strategic business decisions.12
Predictive analytics powered by AI allows businesses to forecast sales, optimize inventory levels, and enhance
supply chain management.13 Customer service AI chatbots provide instant responses to customer inquiries and
streamline support processes.14 In manufacturing, AI systems optimize production processes and predictive
maintenance, and can improve product development, safety, and supply chain logistics.15
In healthcare, medical professionals utilize AI technologies for diagnostic purposes, with algorithms trained
to analyze medical imaging, such as X-rays and MRIs, to identify conditions like tumors and fractures.16
Additionally, AI use in drug research and discovery can assist researchers to analyze biological data and predict
how different compounds might affect health.17

6
In the automotive industry, new vehicles increasingly feature integrated AI functionality, including advanced
driver-assistance systems (ADAS) such as lane-keeping assist and adaptive cruise control.18 AI is also essential
to the development of autonomous vehicles. These vehicles rely on AI systems to process data from sensors,
cameras, and radar, enabling them to navigate complex environments, avoid obstacles, and make real-time
driving decisions.19
Tools powered by generative AI can produce written content, create visual art, and even compose music,
allowing creators, marketers, and businesses to automate content production.20 AI-powered tools and platforms
designed for business applications continue to proliferate rapidly. According to a recent study by McKinsey,
adoption of AI by business jumped from 50% in 2022 to 72% in early 2024, with most of the growth in
professional services.21
Policy Considerations
As the impact of AI increases, policymakers around the world are considering appropriate regulations to balance
further innovation with the need to protect people from harmful unintended consequences. Public policy
concerns include lack of transparency requirements under current law for training data22 as well as outputs
and outcomes,23 intellectual property protections for work created by humans and used as training data for
algorithms. 24 The use of large databases to train AI models raises concerns over data privacy25 and perpetuation
of systemic biases or inequities present in the training data and broader institutions or systems.26
Notably, AI systems are increasingly used to make consequential decisions such as hiring,27 employment
evaluation,28 healthcare treatment,29 credit screening for financial products,30 and private and public benefit
allocation.31 AI systems often operate as “black boxes,” making their decision-making processes hard to discern,
raising concerns over the inability of users and regulators to understand how consequential decisions are made
and how errors can be remedied.32
When applied to decisions such as review of resumes for job candidates and credit applications, AI models
can perpetuate discrimination and inequity.33 Algorithmic bias is a term for when bias built into an algorithm,
including AI systems, intentionally or unintentionally, results in unfairness, including exclusionary experiences
and discriminatory practices.34 Famously, Dr. Joy Buolamwini’s research found that facial recognition technology
performed differently across different demographic groups, with much higher error rates for women with
darker-colored skin because the algorithm was not trained on a broad range of skin tones and facial structures.35
In a different context, when AI systems are used by financial institutions for lending decisions, there is a risk
that biases in the historical data, as a result of discriminatory policies including redlining, will be replicated,
reinforcing existing disparities.36

7
GenAI can create deceptively realistic media of people or events that did not occur, referred to as deepfakes.37
Deepfakes can be disseminated rapidly through social media and other networks, contributing to the
misinformation and disinformation landscape on platforms that impact the public. Deepfakes can be used by
scammers to defraud people, and by bad actors seeking to influence behavior, including voting by providing
false or misleading information about candidates and how to vote.38 Deepfakes depicting political candidates
or influential individuals saying or doing something that did not occur39 have been shared widely across social
media, including by candidates themselves.40
In addition to deepfakes, GenAI can create hyper-realistic sexually explicit images of individuals, including
generation of child sexual abuse material (CSAM) by predators. Individuals depicted in these images may or
may not be identifiable. There are reports of “sextortion” where these images are generated to cause victims
emotional damage or exchange of a commodity to prevent the release of the deepfake.41 The creation of these
images increases strain on law enforcement to respond to a rapidly increasing volume of abusive material.42
A significant increase in the demand for energy accompanies the growth of AI, slowing the transition from
fossil fuels to renewable energy.43 In 2022, data centers, AI, and cryptocurrency made up about 2% of global
electricity demand.44 Several large technology companies questioned their ability to meet their stated
environmental goals regarding clean energy due to their increased need for energy to power AI services.45 At
the same time, AI tools can improve our understanding of climate change and ability to predict significant
and damaging weather events, improving the efficiency of waste management, 46 and improve the energy
efficiency of the technology itself.47
State Legislation
In Washington, Governor Inslee signed Executive Order (EO) 24-01: Artificial Intelligence on January 30, 2024,
directing state agencies to examine several facets of generative AI use for state government. The EO requires
state agencies to identify generative AI initiatives pertinent to agency operations and objectives, create guidelines
for public sector procurement, uses, and ongoing monitoring of the use of generative AI, develop a training
plan for state workers on generative AI, and propose guidelines for agencies to analyze the impact that adopting
generative AI may have on vulnerable communities and criteria to evaluate equitable outcomes in deployment
and implementation of high-risk use cases.48 Agencies have released several of the EO’s deliverables already, and
the final reports and guidelines requested by the EO are due in January 2025.
In addition, Washington legislators passed a bill establishing the AI Task Force and, separately, a bill applying
the same criminal penalties for sharing, selling, possessing, or viewing depictions of a minor engaged in
sexually explicit conduct to include fabricated images of an identifiable minor.49
Numerous other state Legislatures
also passed bills in 2023 and 2024
to address AI concerns. Legislation
enacted in California, Colorado,
Oklahoma, and Utah garnered
significant attention. The various
bills addressed AI education,
disclosure requirements for use
of GenAI, criminal penalties for
synthetic CSAM, 50 and limitations
on the use of automated decision-
making in specific contexts.
California enacted eighteen bills
to set parameters around artificial
intelligence.51 Additional details
about legislation in other states
is available in Appendix 1 of this
report.

Gov. Jay Inslee signs the AI Task Force bill into law.

8
Federal
On October 4, 2022, the White House released the Blueprint for an AI Bill of Rights, 52 which enumerate
principles to protect AI users from potential harm. These principles are meant to be incorporated in
the design, deployment, and use phases of automated systems. Notably, the AI Bill of Rights includes
protections against algorithmic discrimination. The Bill of Rights states that automated AI systems should
not discriminate based on protected statuses and should be designed and used in an equitable manner,
recognizing that if not designed carefully and thoughtfully, AI systems and models can unintentionally harm
marginalized populations.
On January 26, 2023, the National Institute of Standards and Technology (NIST), part of the U.S. Department
of Commerce that provides a technology, measurement and standards infrastructure to promote U.S.
innovation and industrial competitiveness, 53 released the AI Risk Management Framework (RMF).54 The AI
RMF is an approach to integrate security and privacy in the life cycle of an AI system. Collaboration from the
private and public sector was tantamount to developing the RMF and upholding human rights. However, the
AI RMF is not a requirement and is intended for voluntary use.
In October 2023, President Biden signed Executive Order (EO) 14110 on the Safe, Secure, and Trustworthy
Development and Use of Artificial Intelligence.55 EO 14110 sets forth a pathway for the responsible development
and deployment of AI. EO 14110 emphasizes the technology must not harm or infringe on the privacy and civil
liberties of individuals. Importantly, EO 14110 discusses managing the risks of AI systems particularly within
internal government use.
European Union AI Act
In July 2024, the European Union’s Artificial Intelligence Act (EU AI Act) was enacted, establishing the first
significant legal framework for the regulation on AI systems across the EU. The EU AI Act includes varying
regulations based on four different classifications of risk: unacceptable risk, high risk, limited risk, and minimal
risk.56 The EU Parliament seeks to have human oversight embedded into AI systems, instead of solely relying on
automation. Systems that are classified as unacceptable risk, such as systems used to influence peoples’ behavior
in ways that will or can cause physical or psychological harm or systems that exploit vulnerabilities based on
age, disability, or social or economic status, are banned.57 The EU determined these systems are unacceptable
because they are a threat to people broadly or to specific vulnerable populations. Other systems that are classified
as high-risk, due to the risk analysis of operational function, are assessed before deployment and continuously
throughout their lifecycle. Importantly, individual users will have the ability to file complaints to regulatory
bodies as an additional protection to the infringement of individual rights by these systems. All requirements of
the Act will be effective in July 2026.

9
Task Force Work
The authorizing statute requires the Task Force to issue findings and recommendations on a broad range of
topics:
• A literature review of public policy issues with artificial intelligence, including benefits and risks to
the public broadly, historically excluded communities, and other identifiable groups, racial equity
considerations, workforce impacts, and ethical concerns;
• A review of existing protections under state and federal law for individual data and privacy rights,
safety, civil rights, and intellectual property rights, and how federal, state, and local laws relating to
artificial intelligence align, differ, conflict, and interact across levels of government;
• A recommended set of guiding principles for artificial intelligence use informed by standards
established by relevant bodies, including recommending a definition for ethical artificial intelligence
and guiding principles;
• Identification of high-risk uses of artificial intelligence, including those that may negatively affect
safety or fundamental rights;
• Opportunities to support and promote the innovation of artificial intelligence technologies through
grants and incentives;
• Recommendations on appropriate uses of and limitations on the use of artificial intelligence by state
and local governments and the private sector;
• Recommendations relating to the appropriate and legal use of training data;
• Algorithmic discrimination issues which may occur when artificial intelligence systems are used and
contribute to unjustified differential treatment or impacts disfavoring people on the basis of race,
color, national origin, citizen or immigration status, families with children, creed, religious belief
or affiliation, sex, marital status, the presence of any sensory, mental, or physical disability, age,
honorably discharged veteran or military status, sexual orientation, gender expression or gender
identity, or any other protected class under RCW 49.60.010 and recommendations to mitigate and
protect against algorithmic discrimination;
• Recommendations on minimizing unlawful discriminatory or biased outputs or applications;
• Recommendations on prioritizing transparency so that the behavior and functional components of
artificial intelligence can be understood in order to enable the identification of performance issues,
safety and privacy concerns, biases, exclusionary practices, and unintended outcomes;
• Racial equity issues posed by artificial intelligence systems and ways to mitigate the concerns to build
equity into the systems;
• Civil liberties issues posed by artificial intelligence systems and civil rights and civil liberties
protections to be incorporated into artificial intelligence systems;
• Recommendations as to how the state should educate the public on the development and use
of artificial intelligence, including information about data privacy and security, data collection
and retention practices, use of individual data in machine learning, and intellectual property
considerations regarding generative artificial intelligence;
• A review of protections of personhood, including replicas of voice or likeness, in typical contract
structures, and a review of artificial intelligence tools used to support employment decisions;

10
• Proposed state guidelines for the use of artificial intelligence to inform the development, deployment,
and use of artificial intelligence systems to:
o Retain appropriate human agency and oversight;
o Be subject to internal and external security testing of systems before public release for high-risk
artificial intelligence systems;
o Protect data privacy and security;
o Promote appropriate transparency for consumers when they interact with artificial intelligence
systems or products created by artificial intelligence; and
o Ensure accountability, considering oversight, impact assessment, auditability, and due
diligence mechanisms;
• A review of existing civil and criminal remedies for addressing potential harms resulting from the use
of artificial intelligence systems and recommendations, if needed, for new means of enforcement and
remedies; and
• Recommendations for establishing an ongoing committee that must study emerging technologies not
limited to artificial technology.

11
Task Force Activities
On March 18, 2024, Governor Inslee signed Engrossed Second Substitute Senate Bill 5838, establishing the
Washington State Artificial Intelligence Task Force administered by the Attorney General’s Office (AGO). The
statute directs the Task Force to “examine the development and use of artificial intelligence by private and public
sector entities and make recommendations to the Legislature regarding guidelines and potential legislation for
the use and regulation of artificial intelligence systems to protect Washingtonians’ safety, privacy, and civil and
intellectual property rights.”58 The statute calls for eight subcommittees to address AI issues relating to education
and workforce development, public safety and ethics, health care and accessibility, labor, government and
public sector efficiency, state security and cybersecurity, consumer protection and privacy, and industry and
innovation. The Task Force must deliver findings and recommendations on a broad range of legal and policy
issues specified in the statute in three reports: a preliminary report due by December 31, 2024; an interim report
due by December 1, 2025, and a final report due by July 1, 2026.
The Task Force consists of nineteen members representing specific stakeholders and impacted communities
identified by the Legislature. The President of the Senate and the Speaker of the House each appointed one
member from the two largest caucuses of the Senate and House of Representatives, respectively. The Attorney
General appointed the other 15 members representing a broad cross-section of constituencies, including
community and business groups.59 Attorney General Ferguson announced Task Force appointments on June
21, 2024.60
The Task Force held its first meeting on Friday, July 26 (Minutes from this meeting are available on the
AGO website at https://www.atg.wa.gov/aitaskforce). During this meeting, Task Force members took public
comments, reviewed proposed bylaws, and discussed the formation of a leadership committee to direct the
Task Force’s activities.61 Task Force members reviewed the subcommittee structure and discussed changes. After
discussion, Task Force members agreed to add three subcommittees to the eight named in the statute: (1) Ethical
& Responsible AI/AI Governance; (2) Transportation; and (3) Climate & Energy. In addition, the Task Force
proposed renaming the Labor subcommittee to Labor & Employment, to clarify that the subcommittee intended
to focus on issues related to deployment of AI in the workplace in addition to issues pertinent to organized labor.
The Task Force concurred on proceeding with the following eleven subcommittees:
1. Education and Employment;
2. Public Safety;
3. Health Care & Accessibility;
4. Labor and Employment
5. Government and Public Sector Efficiency;
6. State Security and Cybersecurity;
7. Consumer Protection and Privacy;
8. Industry and Innovation;
9. Climate and Energy;
10. Transportation; and
11. Ethical and Responsible AI/AI Governance.
In August and September, the Task Force held the first round of subcommittee meetings. These initial meetings
were open to public participation and structured to solicit input on topics of interest and concern for the
Task Force to consider. During the meetings, participants shared their experiences with AI and concerns
and questions regarding its design, deployment, and use. Task Force members discussed possible focus
areas and how best to organize the Task Force and its subcommittees to work effectively to develop policy
recommendations. Minutes and links to recordings62 for subcommittee meetings are available on the AGO
website at: https://www.atg.wa.gov/aitaskforce.
The second quarterly Task Force meeting occurred on Tuesday, October 29. During this meeting, Task Force
members formally approved its governing bylaws and the minutes from the July 26 meeting. The Task Force
then voted to establish a leadership committee consisting of six members that includes bipartisan legislators and
representatives from state government agencies, labor, business, and an at-large member.

12
The elected leadership committee members are set forth below:
• Senator Joe Nguyen
• Representative Travis Couture
• Katy Ruckle, WaTech
• Cherika Carter, Washington State Labor Council
• Kelly Fukai, Washington Technology Industry Association
• Crystal Leatherman, Washington Retail Association
The Task Force then discussed and adopted changes to reorganize and consolidate subcommittees and clarify
Task Force members’ roles and responsibilities regarding subcommittee participation. During the initial
subcommittee meetings, several participants noted that the large number of subcommittees and the breadth
of each could create operational challenges and lead to potentially duplicative or overlapping workstreams.
The Task Force voted to combine the following subcommittees: (1) the Industry and Innovation and Climate
and Energy subcommittees; (2) the Labor and Employment and Transportation subcommittees, and (3) the
Government and Public Sector Efficiency and State Security and Cybersecurity subcommittees. No changes
were made to the remaining subcommittees.
Updated Subcommittees:
1. Education & Workforce;
2. Public Safety;
3. Health Care & Accessibility;
4. Labor & Employment/Transportation;
5. Government & Public Sector Efficiency/State Security & Cybersecurity;
6. Consumer Protection & Privacy;
7. Industry & Innovation/Climate & Energy; and
8. Ethical & Responsible AI/AI Governance.
Task Force members then committed to participating in at least two subcommittees to provide consistency
and promote alignment between each subcommittee and the full Task Force. The Task Force members on each
subcommittee are listed by last name below:

Subcommittee Members (by last name)
Labor & Employment/Transportation Carter, Tamaru, Leatherman
Health Care & Accessibility Sawyer, Tamaru, Balazinska, Ruckle
Public Safety Lowe, Leatherman, Nguyen, Koshiyama
Consumer Protection & Privacy Ruckle, Sannon, Leatherman, Sardinas, Frank, Harkins
Education & Workforce Balazinska, Tamaru, Fukai, Lowe
Ethical AI & AI Governance Koshiyama, Harkins, Sannon, Shavers
Industry & Innovation/Climate & Energy Fukai, Sardinas, Balazinska, Shavers, Nguyen
Government & Public Sector Efficiency/State Security & Sawyer, Frank, Boehnke, Ruckle, Carter
Cybersecurity

Additionally, the Task Force received an update on the activities undertaken by state agencies pursuant to
Governor Inslee’s executive order directing state agencies to develop guidelines on the use of AI by the state,
then discussed legislation related to AI recently proposed and enacted in other states.

The Task Force will hold its next quarterly meeting in January. In the interim, AGO staff will work with
subcommittees to define priorities and launch regular recurring meetings.

13
Engagement
The benefits, implications and impacts of AI, including its development, deployment, and use, extend to a vast
range of entities, communities, and individuals. While the 19 Task Force members bring significant expertise
and experience to bear, it is not possible for any 19 individuals to represent the full range of experiences and
viewpoints needed to develop meaningful and well-considered recommendations.
The AGO staff commits to engaging communities across the state and removing barriers to participation in
the Task Force. Members of the public are invited to participate in subcommittees. Already, many interested
individuals and people representing public agencies, non-profits, and private businesses participated in
subcommittee discussions and provided valuable insights. The AGO team will continue efforts to broaden
engagement in these conversations.
In addition, the AGO created two advisory groups to provide dedicated spaces for important discussions about
the potential benefits and harms of AI for specific groups. The AGO and Task Force may create additional
advisory groups as the Task Force work proceeds.
Business Advisory Group
The Business Advisory Group will provide regular updates to the Task Force on practical considerations,
challenges, opportunities, needs, and concerns for Washington businesses of different sizes and across
industries in developing, deploying, and using AI. In particular, the advisory group will provide information
on current uses and trends of AI across industries, as well as information about how businesses of different
sizes, in different industries, are considering benefits, risks and potential harms and balancing transparency,
accountability, equity, privacy, ethics, civil and intellectual property rights, and innovation.
The advisory group is convened and led by the Association of Washington Business and the National Federation
of Independent Business.
Tribal Advisory Group
The AGO recognizes that the Legislature did not include Tribes in the Task Force membership. Tribes have
particular needs, concerns, and potential applications for AI use that require a dedicated space for discussion
among Tribal leaders to honor Tribal sovereignty.
AGO Tribal Liaison Asa Washines leads engagement for the Tribal Advisory Group, in close collaboration with
AGO AI Task Force staff.

14
Recommendation
The Legislature should expand liability to cover the creation, possession, or sharing of fabricated images
of a minor engaged in sexually explicit conduct, regardless of whether the minor is identifiable.

Background
In 2024, the Legislature unanimously adopted Substitute House Bill (SHB) 1999, sponsored by Representative
Tina Orwall, addressing AI-generated child sexual abuse material (CSAM). The bill expanded existing criminal
offenses prohibiting dealing in, sending, or bringing into Washington, possessing, or viewing depictions of a
minor engaged in sexually explicit conduct to include fabricated depictions of an identifiable minor.63

SHB 1999 uses the term fabricated images to include images created by generative AI in response to a prompt,
and manipulated images, such as when a person’s face is transposed onto an image of another body, depicting
acts that did not happen. Real CSAM images depicting sexual abuse of children are part of the training data that
enables AI models to produce fabricated CSAM. An investigation by the Stanford Internet Observatory in 2023
found hundreds of known CSAM images in an open dataset that is used to train generative AI models that create
images from text.64 These known images were primarily identified by hash values, which are unique identifiers
for an image, similar to fingerprints.65 When an image is changed, even slightly, the hash value changes, and
scanning software that recognizes known hash values will not recognize images with unknown hash values.66
This exacerbates strain on law enforcement resources as they attempt to discern manipulated known CSAM
from CSAM depicting unidentified victims of child sexual abuse.

The National Center for Missing and Exploited Children (NCMEC) CyberTipline is the designated reporting
mechanism for the public and from electronic service providers to report instances of suspected CSAM or
child sexual exploitation.67 In 2023, the CyberTipline received more than 35.9 million reports, including
more than 100 million files, related to incidents of suspected CSAM, an increase of more than 20% over three
years.68 NCMEC’s Child Victim Identification Program (CVIP) receives reports about CSAM and requests for
assistance from law enforcement to analyze unidentified children in CSAM for potential location or abuser
information. In 2023, CVIP received 4,673 requests from law enforcement, containing more than 32 million
images and videos, to assist in identification of victims.69

Unfortunately, the inclusion of “identifiable” in the legislation makes the law more difficult to enforce and is
counter-productive to efforts to protect children. Images that do not depict real instances of child sexual abuse
strain law enforcement resources and make it harder to identify real child victims.70 Public testimony on SHB
1999 also flagged “…procedural challenges in circumstances where law enforcement officers discover an image
but do not know whether it is authentic or fabricated.”71

To ensure that the law fully protects children, the Task Force recommends that the Legislature remove the
“identifiable” language from the statute.

15
Looking Ahead
In 2025, the Task Force will build on its formative work to develop findings and recommendations as directed
by the statute. The Leadership Committee will guide the priorities and workplans for the Task Force, and will
work with each subcommittee to develop policy proposals for the Task Force to consider recommending to the
Legislature and the Governor’s office. The next Task Force report is due December 1, 2025, and the Task Force
expects to make interim recommendations for policymakers to consider throughout 2025.
The initial focus of the Task Force will be to establish a set of principles that define the essential elements
of trustworthy AI. Trustworthy AI principles are crucial for policy development to provide a framework
that helps governments, organizations, and institutions establish rules, regulations, and strategies that
guide the responsible and beneficial development and use of AI, minimize harm, and respect fundamental
rights. Trustworthy AI refers to AI systems that are designed, developed, and deployed in a manner to be
transparent, reliable, ethical, and aligned with human values. This ensures these systems can be trusted by users,
stakeholders, and society to perform tasks in a way that is equitable, accountable, and beneficial.
Transparent AI systems, which are understandable and accessible to users, developers, and regulators, are a
cornerstone of trustworthy AI. Effective policymaking to maximize the benefits of AI and minimize the risks and
harms requires an understanding of how AI systems are developed and when they are used. When AI systems
are transparent, they are understandable, manageable, and accountable. The Task Force will examine policy
recommendations to provide users and regulators appropriate transparency into how AI systems are developed,
trained and deployed to promote user trust and regulatory oversight.
Beginning with the priorities of developing recommendations for AI principles and meaningful transparency
for Task Force consideration in 2025, each subcommittee will meet regularly to develop the findings and
recommendations outlined in the statute. Below are areas that each subcommittee intends to investigate:
Ethical AI/AI Governance:
This subcommittee will focus on developing recommended guiding principles for AI use, informed by standards
established by relevant bodies, to include a definition for ethical AI. Based on the guiding principles, the
subcommittee will develop a proposed AI governance infrastructure to support implementation of the guiding
principles.
Consumer Protection & Privacy:
This subcommittee will focus specifically on transparency, developing recommendations so that the behavior
and functional components of AI systems can be understood to enable the identification of performance
issues, safety, privacy, and equity concerns, biases, exclusionary practices, and unintended outcomes. The
subcommittee will also develop recommendations to address algorithmic discrimination or bias issues in AI
systems.
Public Safety:
This subcommittee will focus on identification of high-risk uses of AI, including those that may negatively affect
public safety or fundamental rights. The subcommittee will examine use of AI by law enforcement and private
entities to develop potential recommendations regarding appropriate parameters for transparency, human
oversight, and other considerations.
Healthcare & Accessibility:
This subcommittee will examine uses of AI to assist in decisions (e.g., diagnostics or treatment
recommendations) and activities (e.g., email communications, visit summaries) related to patient care, the
allocation of public health benefits, and eligibility for insurance coverage. The subcommittee will focus on
issues related to patient safety, provider responsibility, data privacy and security, and algorithmic bias and
discrimination in the provision and administration of patient care.

16
Education & Workforce:
This subcommittee will explore challenges and opportunities for AI in education (both the uses of AI in
education and education about AI) and develop policy recommendations regarding AI in schools, and how best
to provide relevant AI education at the K-12, community college, undergraduate and graduate school levels. The
subcommittee will also examine the impacts of AI on the workforce and how best to enable a technologically
literate workforce.
Industry & Innovation/Climate & Energy:
This subcommittee will explore strategies to support and promote equitable access to opportunities and
infrastructure for innovation, including through grants, incentives and public/private partnerships. The
subcommittee will also examine the impacts of the increased need for energy to develop and deploy AI systems,
and the ability of public and private entities to increase energy efficiency and meet climate goals.
Government & Public Sector Efficiency/State Security & Cybersecurity:
This subcommittee will build on the work of state agencies under Governor Inslee’s Executive Order to provide:
(1) guidance on the potential uses of AI by state agencies; (2) guidelines for public sector procurement, use, and
monitoring of AI services; (3) training for state employees on AI; (4) guidelines for state agencies to analyze
the impact that adopting generative AI technology may have on vulnerable communities, including criteria to
evaluate equitable outcomes in deployment and implementation of high-risk use cases; and (5) assessing the
impact of AI on state security and cybersecurity.
Labor & Employment/Transportation:
This subcommittee will examine how AI impacts workers and employment, including worker protections
and benefits, issues related to job displacement, job quality and worker well-being, opportunities to maximize
augmentation and minimize displacement, and worker education and reskilling to ensure that workers can
adapt to changes driven by AI and automation. As part of this work, the subcommittee will examine the specific
impacts of AI on workers in the transportation sector.

17
Appendix 1: 5838 Full Text

ENGROSSED SECOND SUBSTITUTE SENATE BILL 5838

AS AMENDED BY THE HOUSE
Passed Legislature - 2024 Regular Session
State of Washington 68th Legislature2024 Regular Session
By Senate Ways & Means (originally sponsored by Senators Nguyen, Conway, Dhingra, Frame, Hasegawa, Hunt, Keiser,
Kuderer, Liias, Lovelett, Lovick, Muzzall, Nobles, Saldaña, Salomon, Stanford, Torres, Valdez, and Wellman; by request of
Attorney General)
READ FIRST TIME 02/05/24.

AN ACT Relating to establishing an artificial intelligence task force; creating new sections; providing an expiration date;
and declaring an emergency.

BE IT ENACTED BY THE LEGISLATURE OF THE STATE OF WASHINGTON:

NEW SECTION. Sec. 1. The legislature finds that artificial intelligence is a fast-evolving technology that holds
extraordinary potential and has a myriad of uses for both the public and private sectors. Advances in artificial intelligence
technology have led to programs that are capable of creating text, audio, and media that are difficult to distinguish from
media created by a human. This technology has the potential to provide great benefits to people if used well and to cause
great harm if used irresponsibly.
The legislature further finds that generative artificial intelligence has become widely available to consumers and has great
potential to become a versatile tool for a wide audience. It can streamline tasks, save time and money for users, and
facilitate further innovation. Artificial intelligence has the potential to help solve urgent challenges, while making our
world more prosperous, productive, innovative, and secure when used responsibly.
Washington state is in a unique position to become a center for artificial intelligence and machine learning. When used
irresponsibly, artificial intelligence has the potential to further perpetuate bias and harm to historically excluded groups. It
is vital that the fundamental rights to privacy and freedom from discrimination are properly safeguarded as society
explores this emerging technology.
The federal government has not yet enacted binding regulations, however in July 2023, the federal government
announced voluntary commitments by seven leading artificial intelligence companies, including three companies
headquartered in Washington, to move toward safe, secure, and transparent development of artificial intelligence
technology. The October 2023 executive order on the safe, secure, and trustworthy development and use of artificial
intelligence builds on this work by directing developers of artificial intelligence systems to share their safety test results
for certain highly capable models with the United States government.
Numerous businesses and agencies have developed principles for artificial intelligence. In Washington, Washington
technology solutions (WaTech) developed guiding principles for artificial intelligence use by state agencies. These
principles share common themes: Accountability, transparency, human control, privacy and security, advancing equity,
and promoting innovation and economic development.
The legislature finds that the possible impacts of advancements in generative artificial intelligence for Washingtonians
requires careful consideration in order to mitigate risks and potential harms, while promoting transparency,
accountability, equity, and innovation that drives technological breakthroughs. On January 30, 2024, Governor Inslee
issued Executive Order 24-01 directing WaTech to identify generative artificial intelligence initiatives that could be
implemented in state operations and issue guidelines for public sector procurement and usage.

NEW SECTION. Sec. 2. (1) Subject to the availability of amounts appropriated for this specific purpose, a task force to
assess current uses and trends and make recommendations to the legislature regarding guidelines and potential
p. 2 E2SSB 5838.SL

18
legislation for the use of artificial intelligence systems is established.
(2) The task force is composed of an executive committee consisting of members as provided in this subsection.
(a) The president of the senate shall appoint one member from each of the two largest caucuses of the senate.
(b) The speaker of the house of representatives shall appoint one member from each of the two largest caucuses of the
house of representatives.
(c) The attorney general shall appoint the following members, selecting only individuals with experience in technology
policy:
(i) One member from the office of the governor;
(ii) One member from the office of the attorney general;
(iii) One member from Washington technology solutions;
(iv) One member from the Washington state auditor;
(v) One member representing universities or research institutions that are experts in the design and effect of an
algorithmic system;
(vi) One member representing private technology industry groups;
(vii) One member representing business associations;
(viii) Three members representing community advocate organizations that represent communities that are
disproportionately vulnerable to being harmed by algorithmic bias;
(ix) One member representing the LGBTQ+ community;
(x) One member representing the retail industry;
(xi) One member representing the hospitality industry;
(xii) One member representing statewide labor organizations; and
(xiii) One member representing public safety.
(d) The task force may meet in person or by telephone conference call, videoconference, or other similar
telecommunications method, or a combination of such methods.
(e) The executive committee may convene subcommittees to advise the task force on the recommendations and findings
set out in subsection (4) of this section.
(i) The executive committee shall define the scope of activity and subject matter focus required of the subcommittees
including, but not limited to: Education and workforce development; public safety and ethics; health care and accessibility;
labor; government and public sector efficiency; state security and cybersecurity; consumer protection and privacy; and
industry and innovation.
(ii) Subcommittees and their members may be invited to participate on an ongoing, recurring, or one-time basis.
(iii) The executive committee in collaboration with the attorney general shall appoint members to the subcommittees that
must be comprised of industry participants, subject matter experts, representatives of federally recognized tribes, or
other relevant stakeholders.
(iv) Each subcommittee must contain at least one member possessing relevant industry expertise and at least one
member from an advocacy organization that represents communities that are disproportionately vulnerable to being
harmed by algorithmic bias including, but not limited to: African American; Hispanic American; Native American; Asian
American; Native Hawaiian and Pacific Islander communities; religious minorities; individuals with disabilities; and
other vulnerable communities.
(v) Meeting summaries and reports delivered by the subcommittees to the executive committee must be made available
on the attorney general's website within 30 days of delivery.
(3) The office of the attorney general must administer and provide staff support for the task force. The office of the
attorney general may, when deemed necessary by the task force, retain consultants to provide data analysis, research,
recommendations, training, and other services to the task force for the purposes provided in subsection (4) of this section.
The office of the attorney general may work with the task force to determine appropriate subcommittees as needed.
(4) The executive committee and subcommittees of the task force shall examine the development and use of artificial
intelligence by private and public sector entities and make recommendations to the legislature regarding guidelines and
potential legislation for the use and regulation of artificial intelligence systems to protect Washingtonians' safety, privacy,
and civil and intellectual property rights. The task force findings and recommendations must include:
(a) A literature review of public policy issues with artificial intelligence, including benefits and risks to the public broadly,
historically excluded communities, and other identifiable groups, racial equity considerations, workforce impacts, and
ethical concerns;
(b) A review of existing protections under state and federal law for individual data and privacy rights, safety, civil rights,
and intellectual property rights, and how federal, state, and local laws relating to artificial intelligence align, differ, conflict,
p. 3 E2SSB 5838.SL

19
and interact across levels of government;
(c) A recommended set of guiding principles for artificial intelligence use informed by standards established by relevant
bodies, including recommending a definition for ethical artificial intelligence and guiding principles;
(d) Identification of high-risk uses of artificial intelligence, including those that may negatively affect safety or fundamental
rights;
(e) Opportunities to support and promote the innovation of artificial intelligence technologies through grants and
incentives;
(f) Recommendations on appropriate uses of and limitations on the use of artificial intelligence by state and local
governments and the private sector;
(g) Recommendations relating to the appropriate and legal use of training data;
(h) Algorithmic discrimination issues which may occur when artificial intelligence systems are used and contribute to
unjustified differential treatment or impacts disfavoring people on the basis of race, color, national origin, citizen or
immigration status, families with children, creed, religious belief or affiliation, sex, marital status, the presence of any
sensory, mental, or physical disability, age, honorably discharged veteran or military status, sexual orientation, gender
expression or gender identity, or any other protected class under RCW 49.60.010 and recommendations to mitigate and
protect against algorithmic discrimination;
(i) Recommendations on minimizing unlawful discriminatory or biased outputs or applications;
(j) Recommendations on prioritizing transparency so that the behavior and functional components artificial intelligence
can be understood in order to enable the identification of performance issues, safety and privacy concerns, biases,
exclusionary practices, and unintended outcomes;
(k) Racial equity issues posed by artificial intelligence systems and ways to mitigate the concerns to build equity into the
systems;
(l) Civil liberties issues posed by artificial intelligence systems and civil rights and civil liberties protections to be
incorporated into artificial intelligence systems;
(m) Recommendations as to how the state should educate the public on the development and use of artificial intelligence,
including information about data privacy and security, data collection and retention practices, use of individual data in
machine learning, and intellectual property considerations regarding generative artificial intelligence;
(n) A review of protections of personhood, including replicas of voice or likeness, in typical contract structures, and a
review of artificial intelligence tools used to support employment decisions;
(o) Proposed state guidelines for the use of artificial intelligence to inform the development, deployment, and use of
artificial intelligence systems to:
(i) Retain appropriate human agency and oversight;
(ii) Be subject to internal and external security testing of systems before public release for high-risk artificial intelligence
systems;
(iii) Protect data privacy and security;
(iv) Promote appropriate transparency for consumers when they interact with artificial intelligence systems or products
created by artificial intelligence; and
(v) Ensure accountability, considering oversight, impact assessment, auditability, and due diligence mechanisms;
(p) A review of existing civil and criminal remedies for addressing potential harms resulting from the use of artificial
intelligence systems and recommendations, if needed, for new means of enforcement and remedies; and
(q) Recommendations for establishing an ongoing committee that must study emerging technologies not limited to
artificial technology.
(5) The executive committee of the task force must hold its first meeting within 45 days of final appointments to the task
force and must meet at least twice each year thereafter. The task force must submit reports to the governor and the
appropriate committees of the legislature detailing its findings and recommendations. A preliminary report must be
delivered by December 31, 2024, an interim report by December 1, 2025, and a final report by July 1, 2026. Meeting
summaries must be posted to the website of the attorney general's office within 30 days of any meeting by the task force.
(6) Legislative members of the task force shall be reimbursed for travel expenses in accordance with RCW 44.04.120.
Nonlegislative members are not entitled to be reimbursed for travel expenses if they are elected officials or are
participating on behalf of an employer, governmental entity, or other organization. Any reimbursement for other
nonlegislative members is subject to chapter 43.03 RCW.

p. 4 E2SSB 5838.SL

20
(7) To ensure that the task force has diverse and inclusive representation of those affected by its work, task force
members, including subcommittee members, whose participation in the task force may be hampered by financial hardship
and may be compensated as provided in RCW 43.03.220.
(8) The definitions in this subsection apply throughout this section unless the context clearly requires otherwise.
(a) "Artificial intelligence" means the use of machine learning and related technologies that use data to train statistical
models for the purpose of enabling computer systems to perform tasks normally associated with human intelligence or
perception, such as computer vision, speech or natural language processing, and content generation.
(b) "Generative artificial intelligence" means an artificial intelligence system that generates novel data or content based
on a foundation model.
(c) "Machine learning" means the process by which artificial intelligence is developed using data and algorithms to draw
inferences therefrom to automatically adapt or improve its accuracy without explicit programming.
(d) "Training data" means labeled data that is used to teach artificial intelligence models or machine learning algorithms to
make proper decisions. Training data may include, but is not limited to, annotated text, images, video, or audio.
(9) This section expires June 30, 2027.

NEW SECTION. Sec. 3. This act is necessary for the immediate preservation of the public peace, health, or safety, or
support of the state government and its existing public institutions, and takes effect immediately.

Passed by the Senate March 4, 2024. Passed by the House February 29, 2024. Approved by the Governor March 18, 2024.
Filed in Office of Secretary of State March 19, 2024.
--- END ---

p. 5 E2SSB 5838.SL

21
Appendix 2: Task Force Bylaws

Adopted by the AI Task Force, Tuesday, October 29, 2024

Washington State Artificial Intelligence Task Force Bylaws

Preamble

In order to fulfill the requirements put forth in the enacted language that establishes the
Washington State Artificial Intelligence Task Force, the membership recognizes that there is a
diverse breadth of experience, expertise, and knowledge that is crucial to completing the work in
front of us. Additionally, these diverse backgrounds also come with unique beliefs, experiences,
and barriers. For that reason, the Task Force membership puts forth these bylaws as a way of
holding each other accountable as we encourage meaningful progress.

Whereas the Washington State Artificial Intelligence Task Force (“Task Force”) is authorized by
the Legislature for the purpose of assessing current uses and trends of artificial intelligence and
making recommendations to the Legislature regarding guidelines and potential legislation for the
use of artificial intelligence systems. Now, therefore, the organization and rules of procedure of
the Task Force shall be as follows:

Purpose
The Task Force is directed to examine a broad range of issues regarding artificial intelligence. The
Task Force's findings and recommendations must include:
• A literature review of public policy issues with artificial intelligence, including benefits and
risks to the public broadly, historically excluded communities, and other identifiable
groups, racial equity considerations, workforce impacts, and ethical concerns;
• A review of existing protections under state and federal law for individual data and privacy
rights, safety, civil rights, and intellectual property rights, and how federal, state, and local
laws relating to artificial intelligence align, differ, conflict, and interact across levels of
government;
• A recommended set of guiding principles for artificial intelligence use informed by
standards established by relevant bodies, including recommending a definition for ethical
artificial intelligence and guiding principles;
• Identification of high-risk uses of artificial intelligence, including those that may negatively
affect safety or fundamental rights;
• Opportunities to support and promote the innovation of artificial intelligence technologies
through grants and incentives;
• Recommendations on appropriate uses of and limitations on the use of artificial
intelligence by state and local governments and the private sector;
• Recommendations relating to the appropriate and legal use of training data;
• Algorithmic discrimination issues which may occur when artificial intelligence systems are
used and contribute to unjustified differential treatment or impacts disfavoring people on
the basis of race, color, national origin, citizen or immigration status, families with children,
creed, religious belief or affiliation, sex, marital status, the presence of any sensory, mental,
or physical disability, age, honorably discharged veteran or military status, sexual
orientation, gender expression or gender identity, or any other protected class under RCW

22
Adopted by the AI Task Force, Tuesday, October 29, 2024

49.60.010 and recommendations to mitigate and protect against algorithmic
discrimination;
• Recommendations on minimizing unlawful discriminatory or biased outputs or
applications;
• Recommendations on prioritizing transparency so that the behavior and functional
components artificial intelligence can be understood in order to enable the identification of
performance issues, safety and privacy concerns, biases, exclusionary practices, and
unintended outcomes;
• Racial equity issues posed by artificial intelligence systems and ways to mitigate the
concerns to build equity into the systems;
• Civil liberties issues posed by artificial intelligence systems and civil rights and civil liberties
protections to be incorporated into artificial intelligence systems;
• Recommendations as to how the state should educate the public on the development and
use of artificial intelligence, including information about data privacy and security, data
collection and retention practices, use of individual data in machine learning, and
intellectual property considerations regarding generative artificial intelligence;
• A review of protections of personhood, including replicas of voice or likeness, in typical
contract structures, and a review of artificial intelligence tools used to support
employment decisions;
• Proposed state guidelines for the use of artificial intelligence to inform the development,
deployment, and use of artificial intelligence systems to:
(i) Retain appropriate human agency and oversight;
(ii) Be subject to internal and external security testing of systems before public release for
high-risk artificial intelligence systems;
(iii) Protect data privacy and security;
(iv) Promote appropriate transparency for consumers when they interact with artificial
intelligence systems or products created by artificial intelligence; and
(v) Ensure accountability, considering oversight, impact assessment, auditability, and due
diligence mechanisms;
• A review of existing civil and criminal remedies for addressing potential harms resulting
from the use of artificial intelligence systems and recommendations, if needed, for new
means of enforcement and remedies; and
• Recommendations for establishing an ongoing committee that must study emerging
technologies not limited to artificial technology.

Article I: Task Force Members
Task Force members are expected to abide by the below code of conduct that ensures a
respectful working environment.

Code of Conduct
Task Force members are expected to, at a minimum:
• Work respectfully with other members of the Task Force, subcommittees, and Attorney
General’s Office (AGO) staff;

23
Adopted by the AI Task Force, Tuesday, October 29, 2024

• Attend and participate in all scheduled Task Force meetings, including community
meetings and public comment sessions;
• Not make any statements or act on behalf of the Task Force independently;
• Serve on at least two subcommittees and regularly attend and participate in
subcommittee meetings;
• Respond timely—and by identified deadlines—to communications from the AGO that
relate to the Task Force and require a response;
• Even when representing yourself or your organization, keep AGO staff informed about all
communications, media interviews and appearances, and any other business related to
Task Force work;
• Share relevant information regarding Task Force work with your constituency;
• Not use your position on the Task Force for personal or organizational gain;
• Identify any actual or potential conflicts of interest that arise and recuse yourselves from
participating in any voting or decision making that gives rise to conflict of interests; and
• Be courageous in challenging biases and previously held beliefs;
• Not engage in comments that perpetuate negative stereotypes about race, gender
identity, sexual orientation, mental health status, housing status, political affiliation, or
religious beliefs; and
• Prioritize care of other members, the community, and families.

The Attorney General’s Office will
• Provide administrative and staff assistance to the Task Force and subcommittees;
• Provide the necessary direction and guidance to subcommittees in their work to advise
the Task Force on findings and recommendations;
• Provide oversight and coordination of the members and subcommittees of the Task
Force;
• Oversee the development of three reports to Washington State Legislature as required
by ESSB 5838; and
• Act as the fiscal agent for the Task Force.

Remova l of Members Violations of this code of conduct may be used as due cause to remove
members of the Task Force. Violations of this code of conduct should be reported to the
leadership committee as soon as possible after the incident(s). Removal of members will be
decided by a simple majority vote of appointed Task Force members at the next scheduled
meeting.

Additionally, the AGO reserves the right to remove members for violations of the above code of
conduct following communication with a Task Force member, including written notice, to inform
them that their conduct is in violation of the code of conduct.

Term of Office The expectation for the term of office is two years or the full course of the Task
Force, if extended by the Legislature. Legislative members are appointed by the President of the

24
Adopted by the AI Task Force, Tuesday, October 29, 2024

The leadership committee represents the majority opinions of the Task Force as a whole.

Remova l of Lea dership Committee Members Any Task Force member may make a motion to
consider the removal of a member of the leadership committee from office for due cause at a
regular meeting of the Task Force, provided that at least seven days written notice of such intent
is given to all Task Force members. The motion to consider removal shall be made at the first
meeting following the required notice. The motion to remove shall be voted on and requires a
three-quarter majority vote, by secret ballot, of all appointed Task Force members to pass.

Article III: Meetings

Time a nd Pla ce Regular meetings of the Task Force shall be held at least twice per year. Meetings
shall be held primarily virtually, with opportunities for the task force to meet in person
periodically. All Task Force meetings will have an option to attend in-person at the Attorney
General’s Office in Seattle.

Task Force meetings will meet for the total time allotted. If business is concluded early, members
may motion to end the Task Force meeting early, with a majority vote of Task Force members in
attendance.

Task Force meetings shall include a designated time for public and community comment that
provides ample time for community members to participate.

Task Force meetings are open to the public and will be advertised at least ten business days prior
to the meeting date.

Quorum A simple majority of Task Force members shall constitute a quorum for the transaction
of business and meeting operations.

Telephone a nd Virtua l Meeting s Members may participate in a meeting of the Task Force by
means of a telephone conference or virtual device, which allows all persons participating in the
meeting to hear each other. Participation virtually or by telephone shall be equivalent to
presence in person at the meeting for purposes of determining if a quorum is present.

Article IV: Attendance

The work ahead of the Task Force is both timely and urgent. The scope of work for the Task
Force is broad and requires all members to participate to the fullest extent possible.
Excused Absences An excused absence is defined as the following:
• Illness;
• School Events and Holidays;
• Religious, Cultural, or Tribal Holidays; or
• Emergency Circumstances.

25
Adopted by the AI Task Force, Tuesday, October 29, 2024

To arrange an excused absence, members should contact AGO staff prior to the meeting with
the specific request.

Remova l of Members When any Task Force member has two unexcused or four excused
absences from regularly scheduled meetings within the calendar year, they shall be notified in
writing by the leadership committee before the next meeting that they may be removed from
the Task Force. The Task Force may remove any member who is absent without excuse from
more than three consecutive meetings. Removal of members requires a majority vote of the
Task Force.

In the event of a vacancy, the AGO will work with the appointing entity for the Task Force
member’s position to fill the vacancy.

Appea l Process If a Task Force member desires to continue serving on the Task Force after
receiving notice of removal, the member must submit a written appeal within one week (seven
calendar days) of receipt of such notice to the AGO indicating an intention to continue serving on
the Task Force.
Article V: Subcommittees

The enacting legislation establishes eight subcommittees. 1 Subcommittees may be added,
removed, or amended as needed by a majority vote of the Task Force throughout the course of
the Task Force. Task Force members should participate regularly in at least two subcommittees.

Chairs or co-chairs of the subcommittees will be confirmed by majority vote of the Task Force.
Time, place, and frequency of the subcommittee meetings are set by the chairs or co-chairs of
the subcommittee and must be published on the AGO website. Subcommittee chairs or co-chairs
will set a recurring meeting cadence that aligns with meeting the deliverables of the authorizing
legislation. Logistical support, including but not limited to meeting space or virtual meeting
support, will be provided by the AGO.

In recognition of the expansive scope of the Task Force, and the far-reaching impacts of artificial
intelligence, members of the public may participate in subcommittee meetings to ensure a broad
range of experience and expertise are represented.

Subcommittee participants must follow guidelines as put forth in these bylaws to ensure
respectful and productive meetings. Subcommittee meetings shall include a designated time for
public and community comment that provides ample time for participation.

1
Education and workforce development; Public safety and ethics; Healthcare and accessibility; Labor; Government
and public sector efficiency; State security and cybersecurity; Consumer protection and privacy; and Industry and
innovation.

26
Adopted by the AI Task Force, Tuesday, October 29, 2024

An AGO staff member will attend all subcommittee meetings. Minutes from each meeting will be
posted on the Task Force’s page of the AGO website within one week following the
subcommittee meeting.

Subcommittee chairs or co-chairs will provide verbal or written updates at each Task Force
meeting on any subcommittee activity since the previous Task Force meeting, and written
recommendations for Task Force consideration when the subcommittee chairs and participating
Task Force members advance a recommendation.

Article VI: Decision Making

Whenever possible, Task Force decisions will be made by group consensus. In the event that
consensus is not possible, a vote with a simple majority will suffice. Dissenting opinions will be
noted in meeting minutes and task force reports. On issues not defined by statutory language,
the Task Force shall consult with the Attorney General’s Office for further interpretation. When
consensus cannot be met, the Task Force leadership committee shall determine appropriate
next steps.

Recommenda tions The Task Force is required to make recommendations for reporting to the
Legislature and the Governor. Subcommittees are responsible for developing recommendations
regarding their subject areas during their regularly scheduled subcommittee meetings. These
recommendations will be developed for consideration by the full Task Force. Recommendations
will be put to roll call votes of Task Force members, and recommendations receiving majority
support will be submitted to the Legislature and Governor.
Proxies In limited circumstances, members may vote by proxy. Proxies may attend meetings.
Intentions to have proxies vote or attend meetings must be submitted in writing to the AGO at
AI@atg.wa.gov at least two hours prior to the meeting. Use of proxies for voting or meeting
attendance should be limited and used only when absolutely necessary.

Electronic Voting Committee members may not vote electronically in accordance with the Open
Public Meetings Act RCW 42.30.060(2).

Article VII: Staff

Staff and meeting facilities will be made available to the Task Force as provided in the AGO
budget. While administratively responsible to the AGO, staff shall act in accordance with
directions and positions of the Task Force in carrying out duties in accordance with the proviso.

27
Appendix 3: Task Force Voting Record
Washington State Artificial Intelligence Task Force
December 16th, 2024
Recommendation to the Legislature to expand liability to cover the creation, possession, or sharing fabricated imag-
es of a minor engaged in sexually explicit conduct, regardless of whether the minor is identifiable.

Task Force Member Aye Nay Abstain Absent
Sen. Joe Nguyen, D-White Center X
Sen. Matt Boehnke, R-Kennewick X
Rep. Clyde Shavers, D-Clinton X
Rep. Travis Couture, R-Allyn X
Sheri Sawyer, Governor’s Office X
Yuki Ishizuka, Attorney General’s Office X
Scott Frank, Auditor’s Office X
Katy Ruckle, Washington Technology Solutions X
Dr. Magdalena Balazinska, University of Washington X
Kelly Fukai, Washington Technology Industry Association X
Ryan Harkins, Microsoft X
Dr. Tee Sannon, ACLU-Washington X
Vicky Tamaru, buildJUSTLY X
Paula Sardinas, WA Build Back Black Alliance X
Leah Koshiyama, Salesforce X
Crystal Leatherman, Washington Retail Association X
Vacant, Hospitality industry representative
Cherika Carter, Washington State Labor Council, AFL-CIO X
Chief Darrell Lowe, Redmond Police Department X

The following Task Force members were unable to attend the meeting and participate in the vote but expressed their
support of the recommendation:
• Sheri Sawyer, Governor’s Office
• Scott Frank, Auditor’s Office
• Vicky Tamaru, buildJUSTLY

28
Appendix 4: AI Policies in Other Jurisdictions
Summary of Proposed and Enacted AI Legislation in the US and Abroad 1

Regulating artiﬁcial intelligence (AI) is a critical priority for governments worldwide as AI technologies
are increasingly integrated into diverse sectors, including healthcare, ﬁnance, and national security. In
the United States, the rapid advancement of AI garners signiﬁcant public interest and attention from
policymakers.

While Congress has not enacted binding or comprehensive regulation on the development, use, or
deployment of AI, a series of presidential executive orders (EOs) directed federal agencies to develop
principles, policies and governance frameworks for adopting AI in a safe and trustworthy manner. At the
state level, state legislatures are considering and passing an increasing number of bills to regulate AI
development broadly or its use for speciﬁc purposes. Internationally, the European Union (EU) passed a
comprehensive, risk-based regulatory framework, and several nations including Brazil, China, Canada,
and Singapore are in the process of implementing AI regulations, though their approaches vary greatly.
This document provides an overview of the evolving landscape of AI legislation, focusing on the federal
approach, state-level initiatives, and the international regulatory environment.

Federal Approach to AI Regulation

At the federal level, presidential executive orders guide the national AI strategy and frame the broader
policy direction as the primary method for policymaking regarding AI. While members of Congress
introduced bills addressing AI as early as 2016, only a few advanced, and none led to comprehensive
federal regulatory frameworks.

The Obama Administration. In 2016, the Obama administration introduced the National Artiﬁcial
Intelligence Research and Development Strategic Plan, 2 which set out priorities for federal investments
in AI research and development. The plan emphasized the importance of advancing AI technologies
while ensuring safe and ethical development. It also highlighted the potential of AI to transform
industries like healthcare, education, and criminal justice, and called for the creation of standards to
track the development of high-risk AI systems. The National Science and Technology Council (NSTC),
created under President Obama, further examined opportunities and risks posed by AI. The NSTC
Preparing for the Future of AI 3 report outlined societal implications of AI and urged public-private
collaboration to maximize the beneﬁts of AI technologies. This framework laid the foundation for future
administrations' eﬀorts to regulate and promote AI in ways that balance innovation with safety.

1 The Attorney General’s oﬃce would like to thank the students and faculty of the UW Technology Law & Public Policy

Clinic for their invaluable research assistance in the preparation of this report.
2 “National Artiﬁcial Intelligence Research and Development Strategic Plan,” National Science and Technology Council,

October 2016,
https://obamawhitehouse.archives.gov/sites/default/ﬁles/whitehouse_ﬁles/microsites/ostp/NSTC/national_ai_rd_strate
gic_plan.pdf.
3 “Preparing for the Future of AI,” National Science and Technology Council, October 2016,

https://obamawhitehouse.archives.gov/sites/default/ﬁles/whitehouse_ﬁles/microsites/ostp/NSTC/preparing_for_the_fu
ture_of_ai.pdf.

29
The Trump Administration. In 2019, President Trump signed EO 13859, "Maintaining American
Leadership in Artiﬁcial Intelligence," 4 which aimed to position the U.S. as a global leader in AI. The EO
outlined ﬁve core principles for AI policy, focusing on fostering innovation, creating technical standards,
providing worker training and ensuring that AI systems were deployed safely. The Trump administration
emphasized national security and economic competitiveness, seeking to mitigate risks posed by AI while
encouraging development and innovation.

On January 1, 2021, Congress passed the National Artiﬁcial Intelligence Initiative Act of 2020, 5 which
established the National AI Initiative Oﬃce, tasked with overseeing the implementation of a national AI
strategy and ensuring that federal agencies collaborate to promote AI research and innovation.

The Biden Administration. Responding to increased public interest in AI, the Biden administration
introduced a more comprehensive approach to AI regulation. In October 2022, President Biden’s Oﬃce
of Science and Technology Policy (OSTP) released an AI Bill of Rights, 6 which emphasized the need for AI
systems to be safe, non-discriminatory, and transparent. The Bill of Rights highlighted ﬁve core
principles: (1) safe and eﬀective systems, (2) algorithmic discrimination protections, (3) data privacy, (4)
notice and explanation, and (5) human alternatives and oversight.

In January 2023, the National Institute of Standards and Technology (NIST), part of the U.S. Department
of Commerce, released the Artiﬁcial Intelligence Risk Management Framework (AI RMF 1.0), 7 a voluntary
framework designed to guide organizations in the responsible development of AI. The framework
provides a structured approach for organizations to assess, manage, and mitigate AI-related risks. The
framework emphasizes strong governance and risk management throughout the entire AI lifecycle, from
design to deployment and monitoring. It focuses on identifying various risks associated with AI systems,
including technical, ethical, legal, and societal concerns like algorithmic bias, fairness, transparency,
safety, and privacy. The framework provides strategies for mitigating these risks, such as using testing
methods like red team testing, and ensuring systems are transparent and auditable. It also advocates for
continuous monitoring of AI systems post-deployment to assess their impacts on users and society, and
encourages involving diverse stakeholders, including developers, users, and policymakers, in the process.

In October 2023, President Biden issued EO 14110, "Safe, Secure, and Trustworthy Development and Use
of Artiﬁcial Intelligence," 8 which built on the principles outlined in the AI Bill of Rights. This EO set the
Administration’s policy direction on the responsible development and use of AI in the United States,
emphasizing innovation, security, and ethical considerations. The order required developers of the most

4
“Executive Order on Maintaining American Leadership in Artiﬁcial Intelligence,” EO 13859, February 2019,
https://trumpwhitehouse.archives.gov/presidential-actions/executive-order-maintaining-american-leadership-artiﬁcial-
intelligence/.
5 P.L. 116-283, National Artiﬁcial Intelligence Initiative Act of 2020, part of the William M. (Mac) Thornberry National

Defense Authorization Act for Fiscal Year 2021, H.R. 6385 Division E, https://www.congress.gov/bill/116th-
congress/house-bill/6395/text.
6
“Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People,” White House Office of
Science and Technology Policy, October 2022, https://www.whitehouse.gov/wp-content/uploads/2022/10/Blueprint-for-
an-AI-Bill-of-Rights.pdf.
7
“Artificial Intelligence Risk Management Framework (AI RMF 1.0),” National Institute of Standards and Technology,
January 2023, https://nvlpubs.nist.gov/nistpubs/ai/NIST.AI.100-1.pdf.
8
“Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artiﬁcial Intelligence,” E.O. 14110,
October 30th, 2023, https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-
the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/.

30
powerful AI systems to share their safety test results and other critical information with the U.S.
government and called on federal agencies to develop policies regarding the use of AI. The EO required
federal agencies to develop standards for content authentication and watermarking to clearly label AI-
generated content; establish cybersecurity standards; develop privacy protections and data privacy
policies; establish guidance to prevent algorithmic bias; develop programs to address worker impacts;
and provide guidelines to federal agencies on the procurement of AI services.

Proposed Federal Legislation
While members of Congress introduced several bills related to AI, Congress has not yet enacted any of
these bills. Below are some of the most signiﬁcant federal legislative proposals to date.

Election Integrity. The rise of deepfakes, synthetic media, and AI-powered tools to manipulate
perceptions and spread misinformation raises serious concerns about election integrity. In response,
Senator Amy Klobuchar introduced three bills to address AI’s impact on elections: the AI Transparency in
Elections Act of 2024, S. 3875, the Protect Elections from Deceptive AI Act, S. 2770, and the Preparing
Election Administrators for AI Act, S. 3897. 9 These proposed laws, while diﬀering in scope and approach,
share the goal of enhancing transparency in the use of AI in political campaigns and protecting the public
from misleading or malicious uses of AI-generated content.

The three bills primarily focus on regulating AI-generated content in political advertisements and
combating deceptive practices during election cycles. One of the central features of these proposed laws
is the requirement for warning labels on materially altered or synthetically generated political content to
help voters distinguish between authentic and AI-manipulated content, mitigating the risk of
misinformation that can inﬂuence public opinion and undermine the democratic process. These bills
address concerns about AI's potential to create highly realistic fake videos, audio recordings, and images
that can deceive voters.

Deepfakes and National Security. National security concerns related to AI-generated content also
prompted federal legislation. The Deepfake Report Act of 2019 10 highlights the potential threats posed
by AI-driven “digital content forgery.” This bill speciﬁcally addresses deepfakes—manipulated media that
can fabricate visual or audio content to deceive viewers. Aspects of the bill were incorporated into the
National Defense Authorization Act (NDAA) for Fiscal Year 2020, making it the ﬁrst federal law to
speciﬁcally address the issue of deepfakes in the context of national security. 11

The law mandates the Director of National Intelligence (DNI) to compile comprehensive assessments
regarding the potential weaponization of deepfake technology by foreign adversaries. 12 These

9 Harper, T. and Branum, B., “Senate Rules Committee Advances Bills to Address Harmful AI in Elections,” Center for

Democracy and Technology, June 6, 2024 https://cdt.org/insights/senate-rules-committee-advances-bills-to-address-
harmful-ai-in-elections/;AI Transparency in Elections Act, S. 3875, 118th Congress, https://www.congress.gov/bill/118th-
congress/senate-bill/3875?q=%7B%22search%22%3A%22ai+elections%22%7D&s=1&r=4; S. Protect Elections from
Deceptive AI Act, 2770, 118th Congress, https://www.congress.gov/bill/118th-congress/senate-
bill/2770?q=%7B%22search%22%3A%22ai+elections%22%7D&s=1&r=1; S. 3897, Preparing Election Administrators for AI
Act, 118th Congress, https://www.congress.gov/bill/118th-congress/senate-
bill/3897?q=%7B%22search%22%3A%22ai+elections%22%7D&s=1&r=3.
10 Deepfake Report Act of 2019, S. 2065, 116th Congress, https://www.congress.gov/bill/116th-congress/senate-

bill/2065/text.
11 Public Law No. 116-92 (December 20, 2019) Sec. 5709.
12 Public Law No. 116-92 (December 20, 2019) Sec. 5709 (a)(2).

31
assessments are intended to evaluate the risks posed by AI-generated disinformation campaigns,
particularly in the context of U.S. elections and national security. The law also requires the DNI to notify
Congress if deepfake activity is identiﬁed as a threat to the integrity of elections.

Transparency. The Algorithmic Accountability Act, H.R. 5628/S. 2892 13 seeks to enhance transparency
and accountability in automated decision-making systems, particularly those inﬂuencing critical aspects
of individuals' lives such as healthcare, housing, education, and employment. If Congress enacts the bill,
it will require companies that deploy automated decision systems to conduct impact assessments to
evaluate the eﬀectiveness and potential biases of these systems. It will also require the creation of a
public repository of algorithms, overseen by the Federal Trade Commission (FTC), to ensure compliance
and facilitate enforcement of the requirements of the act. If passed, this bill would protect consumers
from ﬂawed or biased AI systems and empower them to make informed choices when interacting with
automated decision systems. Members of both chambers of Congress introduced the bill in multiple
sessions, reﬂecting interest in addressing the challenges posed by AI in decision-making processes.

Similarly, the Responsible AI Disclosure Act of 2024, 14 introduced by Representative Maxine Waters on
November 26, 2024, directs ﬁnancial regulatory agencies, including the Federal Reserve and the
Consumer Financial Protection Bureau, to study how companies categorize AI data and train AI models.
The goal is to improve transparency in AI systems within the ﬁnancial sector.

Copyright. The Generative AI Copyright Disclosure Act, H.R. 7913, 15 introduced by Representative Adam
Schiﬀ on April 9, 2024, mandates that companies disclose any copyrighted works used to train generative
AI models at least 30 days before releasing a new or updated AI model. The bill aims to enhance
transparency between AI developers and content creators.

AI Innovation. The Future of Artiﬁcial Intelligence Innovation Act of 2024, S. 4178, 16 introduced by
Senator Maria Cantwell with bipartisan support, focuses on advancing AI research and development to
maintain U.S. competitiveness. It emphasizes support for AI education and workforce training programs.
The Senate Committee on Commerce, Science, and Transportation passed the bill on July 31, 2024, but
the Senate has not taken further action.

Trends in Proposed and Enacted AI Legislation at the State Level

Across the country, state legislatures consider an increasing number of AI-related bills each year. In 2024,
legislators in at least 45 states, Puerto Rico, the Virgin Islands and Washington, D.C., introduced AI bills,
and 31 states, Puerto Rico and the Virgin Islands adopted resolutions or enacted legislation. 17 In total,
state legislatures considered 693 pieces of AI legislation in 2024, in contrast to the 191 pieces of

13 Algorithmic Accountability Act of 2023, H.R. 5628, 118th Congress, https://www.congress.gov/bill/118th-

congress/house-bill/5628?q=%7B%22search%22%3A%22algorithmic+accountability%22%7D&s=2&r=6.
14 Responsible AI Disclosure Act of 2024, H.R. 10263, 118th Congress, https://www.congress.gov/bill/118th-

congress/house-bill/10263?s=7&r=4; as draft legislation https://democrats-
ﬁnancialservices.house.gov/uploadedﬁles/11.26.2024_ai_label_xml.pdf.
15 Generative AI Copyright Disclosure Act, H.R. 7913, 118th Congress, https://www.congress.gov/bill/118th-

congress/house-bill/7913/text.
16 Future of Artiﬁcial Intelligence Innovation Act of 2024, S. 4178, 118th Congress, https://www.congress.gov/bill/118th-

congress/senate-bill/4178.
17 “Summary: Artiﬁcial Intelligence 2024 Legislation,” National Conference of State Legislatures, September 9, 2024,

https://www.ncsl.org/technology-and-communication/artiﬁcial-intelligence-2024-legislation.

32
legislation considered in 2023. 18 Below is an overview of notable proposed and enacted state legislation
on AI.

Regulation of AI Development. Colorado enacted the Colorado Artiﬁcial Intelligence Act, SB 205, in May
2024, becoming the ﬁrst state to enact a comprehensive AI regulatory framework. 19 This bill speciﬁcally
applies to developers and deployers of high-risk AI systems and imposes a duty of reasonable care on
developers and deployers to mitigate risks of algorithmic discrimination.

SB 205 deﬁnes high-risk AI systems broadly as “any artiﬁcial intelligence system that, when deployed,
makes or is a substantial factor in making, a consequential decision.” 20 Developers of high-risk AI systems
must comply with several requirements, including providing detailed information about the AI system,
disclosing any foreseeable risks of algorithmic discrimination to the state attorney general and relevant
deployers, and implementing a risk management policy. Developers also must conduct impact
assessments, review their systems annually for potential discrimination, and provide consumers with
opportunities to correct inaccurate data or appeal consequential decisions made by the AI system. 21

Deployers of high-risk systems also must meet speciﬁc requirements. Among other things, deployers
must implement a risk management policy, complete and update impact assessments, provide
consumers with a statement detailing the role AI played in reaching the consequential decision, and
report any detected algorithmic discrimination to the attorney general within 90 days.

After the Legislature passed the bill, Colorado Governor Jared Polis issued a signing statement,
expressing strong reservations about the bill and his desire for the Legislature to make changes before
the bill takes eﬀect. 22 He speciﬁcally highlighted that the bill “…creates a complex compliance regime for
all developers and deployers doing business in Colorado, with narrow exceptions for small deployers.”
He also expressed concern “…about the impact this law may have on an industry that is fueling critical
technological advancements across our state for consumers and enterprises alike.”

Legislators in several other states introduced similar legislation to SB 205 to require developers to
conduct impact assessments for AI systems. SB 2 in Connecticut was nearly identical, although it faced
signiﬁcant opposition, including a veto threat by the Governor. 23 Lawmakers’ concerns included stiﬂing
innovation and burdening small businesses, as well as making Connecticut an outlier and more diﬃcult
place to do business than other states. 24 Maryland enacted a law requiring impact assessments of AI
systems, though the Maryland law does not list speciﬁc requirements for the impact assessments. 25

18 Curry, H., “2024 State Summary on AI,” Business Software Alliance TechPost, October 22, 2024,

https://techpost.bsa.org/2024/10/22/2024-state-summary-on-ai/.
19 Colorado Artiﬁcial Intelligence Act, S.B. 205, 2024 Leg., Reg. Sess. (Colo. 2024),

https://leg.colorado.gov/sites/default/ﬁles/2024a_205_signed.pdf.
20 Id.
21 Id.
22 Governor Jared Polis, SB 205 Signing Statement, May 17, 2024, https://www.dwt.com/-/media/ﬁles/blogs/artiﬁcial-

intelligence-law-advisor/2024/05/sb24205-signing-
statement.pdf?rev=a902184eafe046cfb615bb047484e11c&hash=213F4C6CDFF52A876011290C24406E7F.
23 S.B. 2, 2024 Leg., Gen. Assem. (Conn. 2024), https://www.cga.ct.gov/2024/TOB/S/PDF/2024SB-00002-R01-SB.PDF.
24 “Sweeping Artiﬁcial Intelligence Bill Stalls in House,” Connecticut Business Industry Association, May 9, 2024,

https://www.cbia.com/news/issues-policies/sweeping-artiﬁcial-intelligence-bill-stalls/
25 Artiﬁcial Intelligence Governance Act of 2024, S.B. 818, 2024 Leg., Reg. Sess. (Md. 2024),

https://mgaleg.maryland.gov/2024RS/Chapters_noln/CH_496_sb0818e.pdf.

33
Proposed bills in Illinois, Massachusetts, Oklahoma, and Virginia included an impact assessment
requirement, though none of these proposed bills moved out of committee. 26

The California Safe and Secure Innovation for Frontier Artiﬁcial Intelligence Models Act, SB 1047, 27 is a
comprehensive AI safety bill that would establish a regulatory framework for the training and
development of advanced AI models. The act speciﬁcally applies to models meeting one of these criteria:
(1) models requiring over 10^26 operations of computing power; or (2) models performing similarly to
state-of-the-art foundation models. It requires several stringent measures for developers of these
models, including conducting comprehensive safety assessments, establishing safety protocols, building
in shut down capabilities, and incident reporting.

Although the Legislature passed SB 1047, California Governor Gavin Newsom vetoed the bill, citing
concerns about its focus on the largest and most expensive AI models. He argued that this focus could
give a false sense of security and potentially hinder beneﬁcial innovations. Additionally, Governor
Newsom noted that the bill did not consider the speciﬁc function or environment of the AI systems,
particularly whether they were used in high-risk contexts or for critical decision-making.

Transparency. In 2024, several states enacted legislation to provide greater transparency into how AI
systems work and when humans interact with AI. In California, AB 2013, enacted in September 2024,
focuses on providing transparency for generative AI training data. 28 The act applies to all developers of
GenAI systems and services, and requires that every time the developers create or substantially modify a
system, they document and post on their website:
(a) The data points contained in the datasets, including sources, owners, types of labels,
total number of data points, intellectual property protection status, and how these
datasets further the intended purpose of the system or device;
(b) Whether the developer purchased or licensed the datasets used by the system or
service;
(c) Whether the datasets include personal information or aggregate consumer information;
(d) The purpose and methodology of any data collection, cleaning, processing, or
modiﬁcation;
(e) The dates on which the datasets were ﬁrst used during development; and
(f) The use of synthetic information for ongoing training and development.

26 Automated Decision Tools Act, H.B. 5116, 103rd. Gen. Assem. (Ill 2024),

https://www.ilga.gov/legislation/fulltext.asp?DocName=&SessionId=112&GA=103&DocTypeId=HB&DocNum=5116&GAI
D=17&LegID=&SpecSess=&Session=; Commercial Algorithmic Impact Assessment Act, H.B. 5322, 103rd Gen. Assem. (Ill
2024).
https://www.ilga.gov/legislation/fulltext.asp?DocName=&SessionId=112&GA=103&DocTypeId=HB&DocNum=5322&GAI
D=17&LegID=&SpecSess=&Session=; Massachusetts Data Privacy Act, H. 4632, 193rd Leg., Reg. Sess. (Mass. 2024),
https://malegislature.gov/Bills/193/H4632; Oklahoma Artiﬁcial Intelligence Bill of Rights, H.B. 3453, 2024 Reg. Sess.
(Okla. 2024), http://www.oklegislature.gov/BillInfo.aspx?Bill=hb3453&Session=2400; Ethical Artiﬁcial Intelligence Act,
H.B. 3835, 2024 Reg. Sess. (Okla. 2024), http://www.oklegislature.gov/BillInfo.aspx?Bill=HB3835&session=2400; Artiﬁcial
Intelligence Developer Act, H.B. 747, 2024 Leg., Reg .Sess. (Va. 2024), https://legacylis.virginia.gov/cgi-
bin/legp604.exe?241+sum+HB747.
27 Safe and Secure Innovation for Frontier Artiﬁcial Intelligence Models Act, S.B. 1047 (Cal. 2024),

https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240SB1047.
28 A.B. 2013, 2024 Leg., Reg. Sess. (Cal. 2024),

https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240AB2013.

34
California and Utah enacted laws that impose disclosure requirements when AI-powered chatbots
interact with humans. Speciﬁcally, California SB 1001 makes it “…unlawful for any person to use a bot to
communicate or interact with another person in California online with the intent to mislead the other
person about its artiﬁcial identity for the purpose of knowingly deceiving the person about the content
of the communication in order to incentivize a purchase or sale of goods or services in a commercial
transaction or to inﬂuence a vote in an election.” 29 The disclosure must be “clear, conspicuous, and
reasonably designed to inform persons with whom the bot communicates or interacts that it is a bot.” 30

Similarly, the Utah AI Policy Act requires that “a person who uses, prompts, or otherwise causes
generative artiﬁcial intelligence to interact with a person in connection with any act administered and
enforced by the Consumer Protection Division of the Utah Department of Commerce . . . shall clearly and
conspicuously disclose to the person with whom the generative artiﬁcial intelligence interacts, if asked or
prompted by the person, that the person is interacting with generative artiﬁcial intelligence and not a
human.” 31

Other states considered but did not enact similar legislation to increase transparency in the use of AI.
Illinois considered but did not pass the Bolstering Online Transparency Act. 32 In Kentucky, SB 266
attempted to require disclosure when AI is used in the purchase or sale of goods or services in a
commercial transaction. 33 The Oklahoma Artiﬁcial Intelligence Bill of Rights, proposed in 2024, would
also have created a similar disclosure requirement. 34

Proﬁling. Indiana, Minnesota, Tennessee, and Virginia all enacted data privacy laws that directly address
the use of AI for proﬁling. 35 The four laws use nearly identical language to deﬁne proﬁling as: “…any
form of solely automated processing performed on personal data to evaluate, analyze, or predict
personal aspects related to an identiﬁed or identiﬁable individual's economic situation, health or health
records, personal preferences, interests, reliability, behavior, location, or movements.“ 36 Under these
laws, consumers have the right to opt out of the processing of their personal data for the purpose of
proﬁling in furtherance of decisions that produce legal or similarly signiﬁcant eﬀects concerning the

29 S.B. 1001, 2024 Leg., Reg. Sess. (Cal. 2024),

https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=201720180SB1001.
30 Id.
31 Utah Artiﬁcial Intelligence Policy Act, S.B. 149, 2024 Leg., Reg. Sess. (Uta. 2024),

https://le.utah.gov/~2024/bills/static/SB0149.html.
32 Bolstering Online Transparency Act, H.B. 5591, 103rd Gen. Assem. (Ill. 2024),

https://www.ilga.gov/legislation/fulltext.asp?DocName=&SessionId=112&GA=103&DocTypeId=HB&DocNum=5591&GAI
D=17&LegID=&SpecSess=&Session=.
33 S.B. 266, 2024 Leg., Reg. Sess. (Ky. 2024), https://apps.legislature.ky.gov/recorddocuments/bill/24RS/sb266/bill.pdf.
34 H.B. 3453, 2024 Leg., Reg. Sess. (Okla. 2024), http://webserver1.lsb.state.ok.us/cf_pdf/2023-

24%20ENGR/hB/HB3453%20ENGR.PDF.
35 S.B. 5, 2024 Leg., Reg. Sess. (Ind. 2024) https://legiscan.com/IN/bill/SB0005/2024; Minnesota Consumer Data Privacy

Act, H.F. 4757, 93rd Leg. Sess. (Minn. 2024),
https://www.revisor.mn.gov/bills/text.php?number=HF4757&type=bill&version=4&session=ls93&session_year=2024&se
ssion_number=0&format=pdf#page=155; Tennessee Information Protection Act, H.B. 1181, 113th Gen. Assemb. (Tenn.
2023), https://publications.tnsosﬁles.com/acts/113/pub/pc0408.pdf; Va. Consumer Data Protection Act, Ch. 53 § 59.1-
580, Code of Virginia, https://law.lis.virginia.gov/vacodefull/title59.1/chapter53/.
36 S.B. 5, 2024 Leg., Reg. Sess. (Ind. 2024) https://legiscan.com/IN/bill/SB0005/2024.

35
consumer. The proposed New York Privacy Act 37 and South Carolina Technology Transparency Act 38
attempted to address proﬁling the same way.

Healthcare. California leads eﬀorts to regulate AI in the healthcare industry. AB 3030, passed in 2024,
requires “a health facility, clinic, physician’s oﬃce, or oﬃce of a group practice that uses generative
artiﬁcial intelligence to generate written or verbal patient communications pertaining to patient clinical
information…ensure that those communications include both (1) a disclaimer that indicates to the
patient that a communication was generated by generative artiﬁcial intelligence, as speciﬁed, and (2)
clear instructions describing how a patient may contact a human health care provider, employee, or
other appropriate person.” 39 SB 1120, also passed in 2024, requires AI usage in the utilization review
and management processes by health care service plans or disability insurers to follow certain
guidelines. 40 These guidelines include ensuring the AI tool bases its determination on speciﬁed
information and is fairly and equitably applied.

Election Misinformation. In 2023, Washington enacted SB 5152 targeting deceptive deepfake use in
election advertisements. SB 5152 creates a cause of action against the sponsor of an electioneering
communication for a candidate whose appearance, action, or speech is intentionally altered through the
use of a synthetic media in the electioneering communication. 41 This prohibition only applies to
communications within the 60 days preceding an election, that clearly identify a candidate for election,
and come from a sponsor whose communications identifying the candidate have a fair market value or
cost of one thousand dollars or more. The depiction must appear to be a real individual, but depict an
appearance, action, or speech which did not occur, and leave a fundamentally diﬀerent understanding or
impression than the original, unaltered version. The law provides for an aﬃrmative defense that the
communication evidently disclosed that the media was manipulated. 42

SB 5152 is generally consistent with election misinformation bills from other states, though there are
some varying provisions. For example, Alabama’s law makes a violation a criminal oﬀense, 43 and

37 New York Privacy Act, S.B. 365B, 2024 Leg., Reg. Sess. (N.Y. 2024),

https://nyassembly.gov/leg/?default_ﬂd=&leg_video=&bn=S00365&term=2023&Summary=Y&Actions=Y&Text=Y.
38 South Carolina Technology Transparency Act, H.B. 4696, 125th Leg., Reg. Sess. (S.C. 2024),

https://www.scstatehouse.gov/sess125_2023-
2024/bills/4696.htm#:~:text=TO%20AMEND%20THE%20SOUTH%20CAROLINA,RIGHTS%2C%20TO%20PROVIDE%20FOR%
20THE.
39 A.B. 3030, 2024 Leg., Reg. Sess. (Cal. 2024),

https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240AB3030.
40 S.B. 1120, 2024 Leg., Reg. Sess. (Cal. 2024),

https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240SB1120.
41 S.B. 5152, 68th Leg., Reg. Sess. (Wash. 2023), https://lawﬁlesext.leg.wa.gov/biennium/2023-

24/Pdf/Bills/Session%20Laws/Senate/5152-S.SL.pdf?q=20241114180141.
42 Id.
43 H.B. 172, 2024 Leg., Reg. Sess. (Ala. 2024),

https://alison.legislature.state.al.us/ﬁles/pdf/SearchableInstruments/2024RS/HB172-enr.pdf.; H.B. 316, 152nd Gen.
Assem., Reg. Sess. (Del. 2024), https://legis.delaware.gov/BillDetail/140956.

36
Hawaii’s statute expands the scope of covered communications. 44 New Hampshire’s law requires
knowledge that the communication was false for liability to attach. 45

Sexually Explicit Material. In 2024, Washington passed HB 1999, which amended existing laws
prohibiting the possession and disclosure of sexually explicit images of minors, to include “fabricated
depictions.” 46 Fabricated depictions are digitally altered or created matters that depict an identiﬁable
individual. 47 HB 1999 is consistent with legislation enacted in other states on the same topic, with two
notable exceptions. For child sexual abuse material, some states do not require that the content be of
an identiﬁable minor. 48 Additionally, some states require that the depiction be so realistic that it is
indistinguishable from a real image or video. 49

Employment Decisions. Legislatures in several states considered legislation regulating employers’ use of
AI systems. Measures requiring employers to notify employees of AI use are most common. State
policymakers are also concerned with the potential for algorithmic discrimination. Notable
considerations include the type of employment decisions covered, and the role AI plays in the ﬁnal
decision.

During Washington’s 2024 legislative session, the Senate Labor and Commerce Committee passed SB
6299, but the bill did not receive a ﬂoor vote. SB 6299 sought to prohibit employers from using AI to
evaluate or otherwise make employment decisions regarding current employees without providing
written disclosure of the employer’s use of AI at an employee’s initial hire or within 30 days of the
employer beginning to use the technology. 50 Notably, this disclosure requirement did not cover hiring. 51

Illinois’s AI Video Interview Act, enacted in 2019, applies to all employers who ask applicants to record
video interviews and use AI tools to analyze those videos. 52 Prior to asking the applicants to submit the
videos, the law requires the employer to notify each applicant that AI may be used, provide each
applicant with information explaining how the AI system works, disclose the characteristics the AI uses to
evaluate applicants, and obtain the consent of each applicant. 53

44 S.B. 2687, 32nd Leg., Reg. Sess. (Haw. 2024),

https://www.capitol.hawaii.gov/session/measure_indiv.aspx?billtype=SB&billnumber=2687&year=2024; H.F. 1370, 93rd
Leg. (Minn. 2024),
https://www.revisor.mn.gov/bills/text.php?number=HF1370&version=latest&session=92&session_number=0&session_y
ear=2023&format=pdf.
45 H.B. 1596, 2024 Leg., Reg. Sess. (N.M. 2024), https://legiscan.com/NH/bill/HB1596/2024.
46 H.B. 1999, 68th Leg., Reg. Sess. (Wash. 2024), https://lawﬁlesext.leg.wa.gov/biennium/2023-

24/Pdf/Bills/Session%20Laws/House/1999-S.SL.pdf?q=20241114170451.
47 Id.
48 H.B. 168, 2024 Leg., Reg. Sess. (Ala. 2024),

https://alison.legislature.state.al.us/ﬁles/pdf/SearchableInstruments/2024RS/HB168-enr.pdf.
49 S.B. 564, 2024 Leg., Reg. Sess. (N.H. 2024),

https://gencourt.state.nh.us/bill_status/legacy/bs2016/billText.aspx?id=2180&txtFormat=html&sy=2024.
50 S.B. 6299, 68th Leg., Reg. Sess. (Wash. 2024), https://lawﬁlesext.leg.wa.gov/biennium/2023-

24/Pdf/Bills/Senate%20Bills/6299-S.pdf?q=20241114155105.
51 Id.
52 Artiﬁcial Intelligence Video Interview Act, H.B. 2557, 101st Gen. Assem. (Ill. 2019),

https://www.ilga.gov/legislation/fulltext.asp?DocName=&SessionId=108&GA=101&DocTypeId=HB&DocNum=2557&GAI
D=15&LegID=&SpecSess=&Session=.
53 Id.

37
In 2024, Illinois enacted HB 3773, requiring employers to notify employees if AI is used for employment-
related decisions.54 HB 3773 also expressly prohibits employer use of AI if it has the eﬀect of unlawful
discrimination with respect to nearly all employment-related decisions, including hiring. 55 According to
the Biden Administration, existing federal antidiscrimination laws already apply to AI algorithmic
discrimination. 56

Job and workforce impacts. Job displacement due to AI is a signiﬁcant concern for policymakers. To
date, most states focus on education and workforce development programs, attempting to ensure that
the economic beneﬁts of AI innovation are spread widely. Maryland’s Talent Innovation Program and
Fund, established by the state Department of Labor pursuant to HB 1128, provides grants for workforce
development programs in the AI, Cybersecurity, Biotechnology, Health Care, and Manufacturing
sectors. 57 In Utah, the Legislature funded the Utah Innovation in Artiﬁcial Intelligence Grant Pilot
Program, which provides grants of up to $1 million for business entities to develop programs, materials,
and curricula for teaching an AI course to students. 58

California recently enacted AB 2602, which voids contractual provisions licensing an individual’s digital
replica rights, where the contract did not contain a reasonably speciﬁc description of the intended use. 59
This bill does not prohibit the use of digital replicas but aims to provide individuals with awareness of,
and the potential opportunity to proﬁt from, digital replicas of their own likeness. Illinois’s Digital Voice
and Likeness Protection Act, passed in 2024, contains very similar provisions. 60 In Washington, the
Legislature in 2024 considered but did not pass SB 6299 to require explicit consent by employees within
the 30 days prior to an employer’s use of the employee’s voice or likeness. 61

AI Development Incentives and Interstate Competition. Several states identify the AI industry as a
potential source of job creation and seek to bring these jobs to their state. Development incentives are
often paired with workforce development or education programs.

In 2024, the New Jersey Legislature created the Next New Jersey Program, a $500 million tax-credit for
businesses primarily engaged in the AI sector, with credits available in amounts up to 25% of capital
invested. 62 Eligibility for this program requires capital investments of $100 million or more in the state,
as well as the creation of 100 or more new full-time jobs.

54 H.B. 3773, 103rd Gen. Assemb. (Ill. 2024),

https://ilga.gov/legislation/fulltext.asp?DocName=&SessionId=112&GA=103&DocTypeId=HB&DocNum=3773&GAID=17
&LegID=&SpecSess=&Session=.
55 Id.
56 “Algorithms, Artiﬁcial Intelligence, and Disability Discrimination in Hiring,” U.S. Department of Justice Civil Rights

Division, May 12, 2022, https://www.ada.gov/resources/ai-guidance/.
57 H.B. 1128, 2024 Leg., Reg. Sess. (Md. 2024),

https://mgaleg.maryland.gov/2024RS/Chapters_noln/CH_421_hb1128e.pdf.
58 S.B. 84, 2024 Leg., Gen. Sess. (Utah 2024), https://le.utah.gov/~2024/bills/static/SB0084.html.
59 A.B. 2602, 2024 Leg., Reg. Sess. (Cal. 2024),

https://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=202320240AB2602.
60 H.B. 4623, 103rd Gen. Assem. (Ill. 2024),

https://www.ilga.gov/legislation/fulltext.asp?DocName=&SessionId=112&GA=103&DocTypeId=HB&DocNum=4623&GAI
D=17&LegID=151978&SpecSess=&Session=.
61 S.B. 6299, 68th Leg., Reg. Sess. (Wash. 2024), https://lawﬁlesext.leg.wa.gov/biennium/2023-

24/Pdf/Bills/Senate%20Bills/6299-S.pdf?q=20241114155105.
62 Next New Jersey Program Act, S.B. 3432, 2024 Leg., Reg. Sess. (N.J. 2024),

https://pub.njleg.state.nj.us/Bills/2024/PL24/49_.PDF.

38
New York passed legislation to provide a $250 million grant to establish the Empire AI Consortium, a
collaborative eﬀort between seven universities to develop AI for social beneﬁt in areas such as climate
change, health care, education, and poverty. 63 This public-private partnership also saw an additional
$125 million invested from the founding institutions and private philanthropic partners.

Maryland’s Pava LaPere Legacy of Innovation Act of 2024 established a $1.5 million grant program for AI
or life science startups in the Baltimore area. 64 Nebraska’s Legislature recently established a grant
program for Nebraska-based businesses researching AI-based writing assistance programs for dyslexia. 65

Government Services/Education. Several states view AI innovation as a tool to provide higher-quality
services to the public.

Hawaii enacted SB 2284 in 2024, requiring the University of Hawaii to develop a wildﬁre forecast system
using AI. 66 Also in 2024, Maryland enacted SB 1068, which directs the Department of Information
Technology to evaluate the feasibility of creating a government services portal using AI, and to prioritize
such a portal’s creation if feasible. 67

Florida and Connecticut both passed laws providing funding for AI use in education. Florida HB 1361
provides grants for school districts to acquire AI tools that provide professional learning to teachers,
lesson plans, insights on student progress, and tutor students. 68 Connecticut HB 5524 funds a pilot
program to provide teachers with AI tools, and training on how to use them, for classroom instruction
and student learning. 69

Trends in Proposed and Enacted AI Regulation Internationally

Internationally, the European Union (EU) leads eﬀorts to regulate AI. Policymakers in other countries are
also establishing AI regulations, which vary from informal guidelines and voluntary frameworks to
binding requirements.

European Union. In April 2021, the European Commission introduced the Artiﬁcial Intelligence Act (AI
Act), 70 the ﬁrst comprehensive legal framework aimed speciﬁcally at governing the development and use
of AI technologies within the EU.

The AI Act uses a risk-management approach, focusing on limiting risks associated with AI systems. The
Act divides AI systems into diﬀerent “risk levels”: Unacceptable, High, Limited, and Minimal risk. The Act

63 S.B. 8308C, 2024 Gen. Assem. (N.Y. 2024), https://www.nysenate.gov/legislation/bills/2023/S8308/amendment/C.
64 Pava LaPere Legacy of Innovation Act of 2024, H.B. 582, 2024 Leg., Reg. Sess. (Md. 2024),

https://mgaleg.maryland.gov/2024RS/Chapters_noln/CH_710_hb0582t.pdf.
65 Leg. B. 1284, 108th Leg. (Neb. 2024), https://nebraskalegislature.gov/FloorDocs/108/PDF/Slip/LB1284.pdf.
66 S.B. 2284, 32nd Leg., Reg. Sess. (Haw. 2024),

https://www.capitol.hawaii.gov/sessions/session2024/bills/SB2284_CD1_.pdf.
67 S.B. 1068, 2024 Leg., Reg. Sess. (Md. 2024), https://legiscan.com/MD/bill/SB1068/2024
68 H.B. 1361, 2024 Leg., Reg. Sess. (Fla. 2024), https://www.ﬂsenate.gov/Session/Bill/2024/1361.
69 H.B. 5524, 2024 Leg., Gen. Assem. (Conn. 2024), https://www.cga.ct.gov/2024/TOB/H/PDF/2024HB-05524-R00-

HB.PDF.
70 Artiﬁcial Intelligence Act, European Union, March 13, 2024, https://www.europarl.europa.eu/doceo/document/TA-9-

2024-0138_EN.html.

39
sets diﬀerent requirements for each risk level, becoming more stringent as the level of risk increases.
Those risks deemed “unacceptable” are banned–such as biometric categorizations, social scoring and
manipulative AI. The majority of the EU AI Act focuses on “high-risk” AI systems to address the most
likely, common, and harmful risks. High risk AI systems must undergo rigorous testing, documentation,
and transparency requirements. Possible high-risk system areas include healthcare, law enforcement,
transportation, and other essential public services. For “limited risk” uses, developers and deployers
must make sure AI users are aware they are interacting with an AI system (such as deepfakes, chatbots,
and other forms of generative AI). Those classiﬁed as “minimal risks” are not regulated under the AI Act.

Brazil. Brazil introduced broad legislation, Bill No. 2,338/2023, regulating AI development based on the
EU risk-management model. This bill would require developers of AI systems to conduct a risk
assessment to identify the level of risk the system may pose before it is placed on the market. The bill
diﬀerentiates between "excessive risk" and "high risk" AI systems. "Excessive risk" AI systems include
those that employ subliminal techniques to aﬀect behavior in ways that are harmful or dangerous to
individuals’ health or safety, exploit the vulnerabilities of groups, and systems that could be used by
governments to evaluate, classify, or rank individuals and attribute a social score. Under the bill’s
framework, a new authority designated by the executive branch will regulate "excessive risk" AI systems.

"High risk" AI systems include systems used in critical infrastructure, professional education and training,
recruitment, employment decisions, evaluation of criteria for accessing public goods, ranking priorities
of emergency response services, administration of justice, autonomous vehicles, and biometric
identiﬁcation systems. The bill mandates that "high risk" AI systems meet quality-of-service
requirements and be designed to allow for human oversight. It mandates regular public impact
assessments for major AI models and systems, holding organizations accountable for their AI
applications, with substantial ﬁnes for non-compliance. 71

Canada. Proposed legislation in Canada also mirrors the EU’s comprehensive risk-based approach. The
Artiﬁcial Intelligence and Data Act (AIDA), 72 introduced in 2022, would establish a harms-based
framework to regulate AI systems, impose substantial ﬁnes for harmful or malicious AI use, and establish
an AI and Data Commissioner to oversee enforcement.

In September 2023, the Canadian government issued a Voluntary Code of Conduct on the Responsible
Management of Advanced Generative AI Systems. This code temporarily provides Canadian companies
with common standards and enables them to voluntarily demonstrate they are developing and using
generative AI systems responsibly until the government enacts formal regulation. 73

China. China regulates AI systems through three diﬀerent laws: Internet Information Service Algorithmic
Recommendation Management Regulations (ARR), Provisions on Management of Deep Synthesis in

71 “Brazil: Introduced Bill No. 2338 of 2023 regulating the use of Artiﬁcial Intelligence, including algorithm design and

technical standards,” Digital Policy Alert, December 10, 2024, https://digitalpolicyalert.org/event/11237-introduced-bill-
no-2338-of-2023-regulating-the-use-of-artiﬁcial-intelligence-including-algorithm-design-and-technical-standards.
72 The Artificial Intelligence and Data Act (AIDA), Bill C-27, Canadian Labour Congress House of Commons’ Standing

Committee on Industry and Technology, November 2023,
https://www.ourcommons.ca/Content/Committee/441/INDU/Brief/BR12711885/br-external/CanadianLabourCongress-
e.pdf.
73 “Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems,”

Government of Canada, September 2023, https://ised-isde.canada.ca/site/ised/en/voluntary-code-conduct-responsible-
development-and-management-advanced-generative-ai-systems.

40
Internet Information Service (Deep Synthesis Regulation or DSR), and Provisional Provisions on
Management of Generative Artiﬁcial Intelligence Services (Generative AI Regulation or GAI). 74

China also takes a risk management approach, specifying various levels of risk, highlighting four levels of
risk similar to the EU AI Act. However, China’s laws focus on how the content is created, released to the
public, and whether it impacts wider People’s Republic of China (PRC) values. The approach to each of
these three considerations diﬀers depending on whether the AI services are accessible or impact
Chinese citizens. AI systems that impact people in the PRC are strictly regulated for content and security
measures, including mandatory risk assessments and ﬁling of AI systems. AI systems created for only
people outside the PRC to use are generally not regulated, and do not require risk assessments.

United Kingdom. The United Kingdom (UK) approach diﬀers from the risk-based regulatory frameworks
by allowing regulators to use current laws to regulate AI within their industries. The UK approach is
based on a report commissioned by the Department of Science. The report, “A Pro-Innovation Approach
to AI Regulation,” is a non-legally binding guidance that outlines a strategic framework for regulating AI,
focusing on key ideas such as adaptability, encouraging innovation, and sector-speciﬁc oversight. 75 This
approach seeks to create an environment of innovation that allows companies and organizations to be
adaptable to the changing landscape of AI evolution.

The UK’s sector-speciﬁc approach allows regulators to interpret the various regulations and laws in their
respective industries as they apply to AI systems. These regulators are tasked with protecting data
privacy and monitoring high-risk applications while allowing for competition and innovation. The UK is
reported to be looking at designing “appropriate legislation” to moderate and control the development
of the most powerful AI models. 76

Singapore. Singapore has taken several steps to promote and regulate AI. The core of Singapore’s
approach is the National AI Strategy, launched in 2019, intended “to develop and use Artiﬁcial
Intelligence (AI) to transform Singapore’s economy and improve the lives of citizens.” 77 This strategy aims
to position Singapore as a global AI hub by integrating AI into critical sectors such as healthcare,
education, ﬁnance, and transportation. 78

74 Toner, H., Creemers, R., and Webster, G., “Translation: Internet Information Service Algorithmic Recommendation

Management Provisions – Effective March 1, 2022,” Stanford University, DigiChina, January 10, 2022,
https://digichina.stanford.edu/work/translation-internet-information-service-algorithmic-recommendation-
management-provisions-effective-march-1-2022/; “Provisions on the Administration of Deep Synthesis Internet
Information Services,” China Law Translate, December 11, 2022, https://www.chinalawtranslate.com/en/deep-
synthesis/#:~:text=Article%206%3A%20Deep%20synthesis%20services,as%20those%20that%20endanger%20the (Dec.
11th, 2022); “Interim Measures for the Management of Generative Artificial Intelligence Services,” China Law Translate,
July 13, 2023, https://www.chinalawtranslate.com/en/generative-ai-interim/.
75 “Policy paper: A pro-innovation approach to AI regulation,” United Kingdom Department for Science, Innovation &

Technology Oﬃce for Artiﬁcial Intelligence, August 3, 2023, https://www.gov.uk/government/publications/ai-regulation-
a-pro-innovation-approach/white-paper.
76 “AI Watch: Global regulatory tracker - United Kingdom,” White & Case, October 18, 2024,

https://www.whitecase.com/insight-our-thinking/ai-watch-global-regulatory-tracker-united-kingdom.
77 “National Artiﬁcial Intelligence Strategy: Advancing Our Smart Nation Journey,” Smart Nation Singapore, November

2019, https://www.smartnation.gov.sg/media-hub/press-releases/national-artiﬁcial-intelligence-strategy-unveiled/.
78 “National AI Strategy: AI for the Public Good, for Singapore and the World,” Smart Nation Singapore, March 2023,

https://ﬁle.go.gov.sg/nais2023.pdf.

41
A key part of Singapore’s regulatory structure is the Model AI Governance Framework. 79 This framework,
much like the NIST AI RMF 1.0, is voluntary and provides a detailed and broad set of guidance to private
sector organizations regarding how to ethically deploy AI systems. It focuses on two foundational
principles: accountability and human-centricity. Organizations are advised to establish robust governance
structures, manage risks, and ensure that their AI operations prioritize fairness, transparency, and the
well-being of individuals. The framework provides real-world examples, an assessment guide, and
alignment with global AI standards such as those set by the EU. 80In addition to the Model AI Governance
Framework, in 2022 Singapore launched AI Verify, a testing toolkit that organizations can use to evaluate
AI systems’ fairness, transparency, and explainability through the standardized testing. 81

Saudi Arabia. Saudi Arabia established the Saudi Data and Artiﬁcial Intelligence Authority (SDAIA) by
royal decree to govern data and AI policy. 82 The SDAIA operates directly under the Prime Minister and is
empowered to develop policies, governance mechanisms, standards, and controls related to data and
artiﬁcial intelligence, and to monitor compliance, to support the Kingdom become an AI leader as part of
its Vision 2030 83 agenda for the economic and social future of the nation.

The SDAIA issued draft AI Ethics Principles in September 2023 (the "AI Principles.") 84 Additionally, the
SDAIA also issued two Generative AI Guidelines (the "Guidelines") on January 1, 2024. The ﬁrst version is
for government employees, 85 and the second is for the public. 86 Both versions provide guidelines
regarding adoption and use of generative AI systems, along with examples based on common scenarios.
The guidelines documents, which do not have the force of law, highlight the challenges and
considerations associated with the use of generative AI, propose principles for responsible use, and
present recommended practices.

79 “Model Artificial Intelligence Governance Framework Second Edition,” Singapore Digital, Infocomm Media

Development Authority and Singapore Personal Data Protection Commission, January 2020, https://www.pdpc.gov.sg/-
/media/Files/PDPC/PDF-Files/Resource-for-Organisation/AI/SGModelAIGovFramework2.pdf.
80 Id.
81 “Invitation to Pilot AI Verify: AI Governance Testing Framework & Toolkit,” Infocomm Media Development Authority,

May 25, 2022, https://file.go.gov.sg/aiverify.pdf; “National AI Strategy: AI for the Public Good, for Singapore and the
World,” Smart Nation Singapore, March 2023, https://ﬁle.go.gov.sg/nais2023.pdf.
82 “About Saudi Data & AI Authority,” Saudi Data & AI Authority, https://sdaia.gov.sa/en/SDAIA/about/Pages/About.aspx.
83 “Saudi Vision 2030,” Kingdom of Saudi Arabia, https://www.vision2030.gov.sa/en.
84 “AI Ethics Principles: Version 1.0” Saudi Data & AI Authority, September 2023,

https://sdaia.gov.sa/en/SDAIA/about/Documents/ai-principles.pdf.
85 “Generative Artiﬁcial Intelligence Guidelines for Government: Version 1,” Saudi Data & AI Authority, January 2024,

https://sdaia.gov.sa/en/SDAIA/about/Files/GenAIGuidelinesForGovernmentENCompressed.pdf.
86 “Generative Artiﬁcial Intelligence Guidelines Public: Version 1,” Saudi Data & AI Authority, January 2024,

https://sdaia.gov.sa/en/SDAIA/about/Files/GenerativeAIPublicEN.pdf.

42
Appendix 5: Election Guidance

STATE OF WASHINGTON
Voter Guide to AI and Elections

Before you cast your ballot, be vigilant: bad actors are increasingly using artificial
intelligence (AI) to spread misinformation and manipulate voters. Protect your vote-
Make sure you have accurate information to guide your choices.

Washingtonians are familiar with scams that attempt to manipulate people into disclosing
financial information. However, during election cycles, bad actors are using advances in
technology, such as AI, to mislead voters in an attempt to manipulate an election. This
document will outline how to identify AI and protect yourself from its misuse.

What is artificial intelligence generated media?

AI is a technology that performs tasks normally associated with human intelligence,
including creating realistic images, videos, and audio.

New and rapidly developing AI tools, broadly called generative AI, can quickly create
high-quality images and media content.

Generative AI can also be used to produce realistic but fake images, videos, and audio,
called synthetic media. This technology is widely accessible, making it easy for bad
actors to create media intended to deceive or defraud others. These bad actors may
attempt to obtain financial information, to influence voting decisions, or even discourage
people voting at all.

For example, images and videos can depict events that never occurred, often referred to
as deepfakes. An audio deepfake manipulates a person’s voice to make false statements,
which can be used in robocalls or spam calls to deceive voters or defraud
Washingtonians.

Generative AI may also be used to spread misinformation and disinformation.
Misinformation and disinformation while used interchangeably have different
meanings. Misinformation is false information that is shared with others, including
inaccurate information shared unintentionally without verification. Disinformation is
deliberately false information intended to deliberately mislead. AI makes the creation and
spread of both misinformation and disinformation easier, and it can cause lasting harm in
the election process.

How can chatbots and search engines impact elections?

Most Washingtonians are familiar with using search engines to quickly find answers
online by typing a question or phrase in a search bar, which quickly yields a wide range
of results. Individuals are now turning to artificial intelligence to ask questions. As more
people begin using AI, such as chatbots, the potential risks associated with
misinformation increases.

43
Chatbots are a type of artificial intelligence that can understand natural language
processing, i.e. human language. Users can ask a question or enter a prompt and the
chatbot generates a personalized response. These AI systems use large volumes of
existing data and user-generated inputs to improve their responses. However, the data can
include errors, leading to incorrect or misleading answers that may be presented as
confirmed facts, this is known as AI hallucinations. Currently, chatbots do not have a
built-in mechanism to verify the accuracy of information provided by these systems. In a
study released in February 2024, chatbots provided incorrect information about voter
registration deadlines and indicated that voters could vote by text message, which is not
legal in any state.i

What can you do to stay informed?
Synthetic content can spread quickly on social media platforms and becomes
uncontrollable once it is released. Distinguishing real content from what is AI-generated
materials is challenging. Currently, there are no reliable mechanisms to identify AI-
generated content or, requirements to disclose AI-generated content. Here are a few
recommendations to protect yourself from misinformation.
o Verify the information you read or share on social media. Technology can
improve the quality of social media content to create realistic posts. The
use of these technologies can increase deceptive content that users believe
is real. Check the original source of the content you share and be critical
of information that does not appear to have an original source or owner. It
is important to know beforehand the information you are sharing to your
social networks is not misleading.
o Verify information about candidates or ballot measures. Cross-reference
information on a candidate’s official websites or their official social media
channels. Verify statements about election procedures and deadlines with
your county election offices or the Washington Secretary of State’s Office
at https://www.sos.wa.gov/elections#voters. Do not rely solely on chatbots
or information produced from an AI engine.
o Report deceptive content on social media platforms. Many social media
platforms allow users to flag misinformation. Be aware of any odd
features in photos and videos, such as distorted features and uneven
lighting and shadows with polarizing messaging, which can be a good
indication of altered or AI-generated media.
o Have a plan to vote. Confirm your voter registration with your local
county election office as well as how to receive or cast your ballot. You
can verify that you are registered to vote and learn whether your ballot has
been mailed, received, or accepted online at votewa.gov. If you see
information online that contradicts your prior knowledge of voting
procedure, such as changes to ballot drop box locations or in-person
voting, speak to your local elections office or the Washington Secretary of
State’s Office at https://www.sos.wa.gov/elections#voters.

i
Picchi, A., “AI chatbots are serving up wildly inaccurate election information, new study says,” CBS
MoneyWatch, February 27, 2024, https://www.cbsnews.com/news/ai-chatbots-inaccurate-election-
information-proof-news/.

44
Endnotes
1 Thorbecke, C. “A year after ChatGPT’s release, the AI revolution is just beginning,” CNN Business, No-
vember 30, 2023. https://www.cnn.com/2023/11/30/tech/chatgpt-openai-revolution-one-year/index.html (Last
accessed November 14, 2024).
2 “Gov. Inslee signs executive order to prepare the state for generative AI,” Washington Technology
Solutions, January 30, 2024, https://watech.wa.gov/watech/news/2024/gov-inslee-signs-executive-order-pre-
pare-state-generative-ai (Last accessed November 14, 2024); “Interim Guidelines for Purposeful and Re-
sponsible Use of Generative Artificial Intelligence,” Washington Technology Solutions, August 8, 2023. https://
watech.wa.gov/sites/default/files/2023-09/State%2520Agency%2520Generative%2520AI%2520Guide-
lines%25208-7-23%2520.pdf (Last accessed November 14, 2024).
3 Siegel, E. “3 Ways Predictive AI Delivers More Value Than Generative AI,” Forbes, March 25, 2024.
https://www.forbes.com/sites/ericsiegel/2024/03/04/3-ways-predictive-ai-delivers-more-value-than-genera-
tive-ai/ (Last accessed November 14, 2024).
4 Johns, S. “How AI is Revolutionizing Social Media Marketing: 5 Innovative Strategies for Instagram
in 2024,” Global Trade, May 16, 2024. https://www.globaltrademag.com/how-ai-is-revolutionizing-social-me-
dia-marketing-5-innovative-strategies-for-instagram-in-2024/ (Last accessed November 26, 2024).
5 Shaota, N. “Streaming Into The Future: How AI Is Reshaping Entertainment,” Forbes, March 18, 2024,
https://www.forbes.com/sites/neilsahota/2024/03/18/streaming-into-the-future-how-ai-is-reshaping-entertain-
ment/ (Last accessed November 26, 2024).
6 Furchgott, R. “The Best Smart Thermostat,” New York Times Wirecutter, August 13, 2024, https://www.
nytimes.com/wirecutter/reviews/the-best-thermostat/ (Last accessed November 26, 2024).
7 Lau, J. “Google Maps 101: How AI helps predict traffic and determine routes,” The Keyword Google Blog,
September 3, 2020, https://blog.google/products/maps/google-maps-101-how-ai-helps-predict-traffic-and-deter-
mine-routes/ (Last accessed November 26, 2024).
8 Antonius, F., et al. “Incorporating Natural Language Processing into Virtual Assistants: An Intelligent
Assessment Strategy for Enhancing Language Comprehension,” International Journal of Advanced Computer Sci-
ence Applications, Vol. 14, No. 10, 2023, p. 741. https://thesai.org/Downloads/Volume14No10/Paper_79-Incorpo-
rating_Natural_Language_Processing.pdf (Last accessed November 13, 2024).
9 Seigel, E. “3 Ways Predictive AI Delivers More Value Than Generative AI,” Forbes, March 4, 2024, https://
www.forbes.com/sites/ericsiegel/2024/03/04/3-ways-predictive-ai-delivers-more-value-than-generative-ai/ (Last
accessed November 26, 2024).
10 Id.
11 Haan, K. “How Businesses Are Using Artificial Intelligence In 2024,” Forbes Advisor, April 24, 2023,
https://www.forbes.com/advisor/business/software/ai-in-business/ (Last accessed November 13, 2024).
12 Madaan, H. and Forbes Technology Council, “How Businesses Can Leverage AI-Powered Deci-
sion-Making For Sustained Growth,” Forbes, October 9, 2024, https://www.forbes.com/councils/forbestechcoun-
cil/2024/10/09/how-businesses-can-leverage-ai-powered-decision-making-for-sustained-growth/ (Last accessed
November 26, 2024).
13 Clarke, H. “Mastering Decision-Making: Machine Learning for Predictive Analytics,” Harrison Clarke,
January 3, 2024, https://www.harrisonclarke.com/blog/mastering-decision-making-machine-learning-for-pre-
dictive-analytics (Last accessed November 26, 2024).

45
14 Pagidyala, S., “Key Lessons From Amazon’s Customer Service Chatbot Q: Innovations in artificial intel-
ligence are continuously refining the quality of customer interactions. Here’s how to make them work for you,”
Inc., February 5, 2024, https://www.inc.com/srini-pagidyala/key-lessons-from-amazons-customer-service-chat-
bot-q.html (Last accessed November 26, 2024).
15 “Working Smarter: How Manufacturers Are Using Artificial Intelligence,” National Association of Manu-
facturers, May 2024. https://nam.org/wp-content/uploads/2024/05/NAM-AI-Whitepaper-2024-1.pdf, p. 4 (Last
accessed November 26, 2024).
16 Umapathy VR, Rajinikanth B S, Samuel Raj RD, Yadav S, Munavarah SA, Anandapandian PA, Mary AV,
Padmavathy K, R A. “Perspective of Artificial Intelligence in Disease Diagnosis: A Review of Current and Future
Endeavours in the Medical Field.” Cureus. September 21, 2023 ;15(9):e45684. doi: 10.7759/cureus.45684. PMID:
37868519; PMCID: PMC10590060. https://pmc.ncbi.nlm.nih.gov/articles/PMC10590060/ (Last accessed No-
vember 26, 2024).
17 Qureshi R, Irfan M, Gondal TM, Khan S, Wu J, Hadi MU, Heymach J, Le X, Yan H, Alam T. AI in drug
discovery and its clinical relevance. Heliyon. July 2023;9(7):e17575. doi: 10.1016/j.heliyon.2023.e17575. Epub
2023 Jun 26. PMID: 37396052; PMCID: PMC10302550. https://pmc.ncbi.nlm.nih.gov/articles/PMC10302550/
(Last accessed November 26, 2024).
18 “How AI is Revolutionizing Automotive Safety: A Boon for Drivers, Passengers, and the Future of Trans-
portation,” Medium, February 22, 2024. https://medium.com/@sahasibaprasad457/how-ai-is-revolutionizing-
automotive-safety-a-boon-for-drivers-passengers-and-the-future-of-2da20ca15233 (Last accessed November 26,
2024).
19 Fisher, J. “Cars with AI in 2024: Volkswagen and Mercedes-Benz Join Tesla at the Forefront,” CarEdge,
May 24, 2024. https://caredge.com/guides/cars-with-ai-in-2024 (Last accessed November 26, 2024).
20 Marr, B. “Generative AI And The Future Of Content Creation,” Forbes, November 30, 2023. https://www.
forbes.com/sites/bernardmarr/2023/11/30/generative-ai-and-the-future-of-content-creation/ (Last accessed No-
vember 26, 2024).
21 The state of AI in early 2024: Gen AI adoption spikes and starts to generate value, McKinsey 2024, the-
state-of-ai-in-early-2024-final.pdf (Last accessed November 26, 2024).
22 Hardings, J, Simperl, E, and Shadbolt, N., “We Must Fix the Lack of Transparency Around the Data Used
to Train Foundation Models,” Harvard Data Science Review, May 31, 2024. https://hdsr.mitpress.mit.edu/pub/
xau9dza3/release/2 (Last accessed November 14, 2024).
23 Leibowicz, C. “Why watermarking AI-generated content won’t guarantee trust online,” MIT Technology
Review, August 9, 2023. https://www.technologyreview.com/2023/08/09/1077516/watermarking-ai-trust-online/
(Last accessed November 14, 2024).
24 Heikkilä, M. “AI companies are finally being forced to cough up for training data,” MIT Technology
Review, July 2, 2024. https://www.technologyreview.com/2024/07/02/1094508/ai-companies-are-finally-being-
forced-to-cough-up-for-training-data/ (Last accessed November 14, 2024).
25 Fazlioglu, M., “Training AI on personal data scraped from the web,” International Association of Privacy
Professionals, November 8, 2023. https://iapp.org/news/a/training-ai-on-personal-data-scraped-from-the-web
(Last accessed November 14, 2024).
26 Schwartz, R., Vassilev, A., Greene, K., Perine, L., Burt, A., and Hall, P. “Toward a Standard for Identifying
and Managing Bias in Artificial Intelligence,” NIST Special Publication 1270, National Institute of Standards and
Technology, March 2022. https://nvlpubs.nist.gov/nistpubs/SpecialPublications/NIST.SP.1270.pdf (Last accessed
November 14, 2024).

46
27 Nelli, S. “Bad Hires Can Cost You Thousands to Replace—and AI May Be to Blame. Try These 3 Ways
to Ensure You Don’t Fall Into Its Trap,” Entrepreneur, November 13, 2024. https://www.entrepreneur.com/sci-
ence-technology/is-over-reliance-on-ai-hiring-making-you-miss-out-on-top/481285 (Last accessed November
14, 2024).
28 McBride, S. “Rippling launches AI-based tool to evaluate employee performance,” Bloomberg
News, September 25, 2024. https://www.bloomberg.com/news/articles/2024-09-25/rippling-introduc-
es-ai-based-tool-to-evaluate-employee-performance (Last accessed November 14, 2024).
29 Collins, B. “New Courses Respond to Rapid Adoption of AI in Health Care,” Harvard Medical School,
October 17, 2024. https://hms.harvard.edu/news/new-courses-respond-rapid-adoption-ai-health-care (Last
accessed November 14, 2024).
30 Gross, P. “As AI takes the helm of decision making, signs of perpetuating historic biases emerge,” Oregon
Capital Chronicle, October 18, 2024. https://oregoncapitalchronicle.com/2024/10/18/as-ai-takes-the-helm-of-de-
cision-making-signs-of-perpetuating-historic-biases-emerge/ (Last accessed November 14, 2024).
31 “Framework for State, Local, Tribal, and Territorial Use of Artificial Intelligence for Public Benefit Ad-
ministration,” U.S. Department of Agriculture, Food and Nutrition Service, April 29, 2024. https://www.fns.usda.
gov/framework-artificial-intelligence-public-benefit (Last accessed November 14, 2024).
32 Blouin, L. “AI’s mysterious ‘black box’ problem, explained,” University of Michigan-Dearborn News,
March 6, 2023. https://umdearborn.edu/news/ais-mysterious-black-box-problem-explained (Last accessed No-
vember 14, 2024).
33 Gross, P. “As AI takes the helm of decision making, signs of perpetuating historic biases emerge,” Oregon
Capital Chronicle, October 18, 2024. https://oregoncapitalchronicle.com/2024/10/18/as-ai-takes-the-helm-of-de-
cision-making-signs-of-perpetuating-historic-biases-emerge/ (Last accessed November 14, 2024).
34 Buolamwini, J. “How I’m fighting bias in algorithms,” TEDxBeaconStreet, November 2019, https://www.
ted.com/talks/joy_buolamwini_how_i_m_fighting_bias_in_algorithms?hasSummary=true&subtitle=en (Last
accessed November 26, 2024).
35 “Unmasking the Bias: How Joy Buolamwini is Fighting for Ethical AI,” Capitology Blog, Capitol Technol-
ogy University, February 19, 2024. https://www.captechu.edu/blog/unmasking-bias-how-joy-buolamwini-fight-
ing-ethical-ai (Last accessed November 26, 2024).
36 Browne, R. and Sigalos, M. “A.I. has a discrimination problem. In banking, the consequences can be
severe.” CNBC, June 23, 2023. https://www.cnbc.com/2023/06/23/ai-has-a-discrimination-problem-in-banking-
that-can-be-devastating.html (Last accessed November 26, 2024).
37 “Deepfake,” Merriam-Webster Dictionary, https://www.merriam-webster.com/dictionary/deepfake (Last
accessed November 14, 2024).
38 “Don’t fall for deepfakes this election,” Microsoft, https://news.microsoft.com/ai-deepfakes-elections/
(Last accessed November 26, 2024).; Dwyer, D., and Herndon, S., “AI deepfakes a top concern for election
officials with voting underway,” ABC News, October 18, 2024. https://abcnews.go.com/Politics/ai-deep-
fakes-top-concern-election-officials-voting-underway/story?id=114202574 (Last accessed November 26, 2024).
39 Verma, P., “A deepfake showed MLK Jr. backing Trump. His daughter calls it ‘vile.’” Washington Post, No-
vember 4, 2024, https://www.washingtonpost.com/technology/2024/11/04/martin-luther-king-jr-trump-deep-
fake/ (Last accessed November 26, 2024).
40 Robins-Early, N., “How did Donald Trump end up posting Taylor Swift deepfakes?” The Guardian,
August 26, 2024, https://www.theguardian.com/technology/article/2024/aug/24/trump-taylor-swift-deepfakes-ai
(Last accessed November 26, 2024).

47
41 “Public Service Announcement: Malicious Actors Manipulating Photos and Videos to Create Explicit
Content and Sextortion Schemes,” Federal Bureau of Investigation, June 5, 2023, https://www.ic3.gov/PSA/Ar-
chive/2023/PSA230605 (Last accessed November 26, 2024).
42 McQue, K., “AI is overpowering efforts to catch child predators, experts warn,” The Guardian, July 18,
2024, https://www.theguardian.com/technology/article/2024/jul/18/ai-generated-images-child-predators (Last
accessed November 26, 2024).
43 Milmo, D., Hern, A., and Ambrose, J. “Can the climate survive the insatiable energy demands of the AI
arms race?” The Guardian, July 4, 2024. https://www.theguardian.com/business/article/2024/jul/04/can-the-cli-
mate-survive-the-insatiable-energy-demands-of-the-ai-arms-race (Last accessed November 14, 2024).
44 Crownhard, C. “AI is an energy hog. This is what it means for climate change.” MIT Technology Review,
May 23, 2024. https://www.technologyreview.com/2024/05/23/1092777/ai-is-an-energy-hog-this-is-what-it-
means-for-climate-change/ (Last accessed November 14, 2024).
45 Milmo, D., Hern, A., and Ambrose, J. “Can the climate survive the insatiable energy demands of the AI
arms race?” The Guardian, July 4, 2024. https://www.theguardian.com/business/article/2024/jul/04/can-the-cli-
mate-survive-the-insatiable-energy-demands-of-the-ai-arms-race (Last accessed November 14, 2024).
46 Masterson, V. “9 ways AI is helping tackle climate change,” World Economic Forum, February 12, 2024.
https://www.weforum.org/stories/2024/02/ai-combat-climate-change/ (Last accessed November 14, 2024).
47 “Reducing AI’s Climate Impact: Everything You Always Wanted to Know but Were Afraid to Ask,” Berke-
ley Gateway to Innovation, September 13, 2024. https://begin.berkeley.edu/reducing-ais-climate-impact-every-
thing-you-always-wanted-to-know-but-were-afraid-to-ask/ (Last accessed November 14, 2024).
48 Executive Order 24-01: Artificial Intelligence, Governor Jay Inslee, January 30, 2024. https://governor.
wa.gov/sites/default/files/exe_order/24-01%20-%20Artificial%20Intelligence%20%28tmp%29_0.pdf (Last ac-
cessed November 14, 2024).
49 Final Bill Report, SHB 1999, Ch. 88, Laws of 2024, Washington State Legislature, https://lawfilesext.leg.
wa.gov/biennium/2023-24/Pdf/Bill%20Reports/House/1999-S%20HBR%20FBR%2024.pdf?q=20241107103458
(Last accessed November 26, 2024).
50 “California Criminalizes AI-Enabled Child Sexual Abuse,” Marc Berman, Assemblymember, Dis-
trict 23, September 29, 2024, https://a23.asmdc.org/press-releases/20240929-california-criminalizes-ai-en-
abled-child-sexual-abuse (Last accessed November 26, 2024).
51 Zeff, M., “Here is what’s illegal under California’s 18 (and counting) new AI laws,” TechCrunch, Sep-
tember 29, 2024, https://techcrunch.com/2024/09/29/here-is-whats-illegal-under-californias-18-and-count-
ing-new-ai-laws/?guccounter=1 (Last accessed November 26, 2024).
52 “Blueprint for an AI Bill of Rights,” The White House Office of Science and Technology Policy, https://
www.whitehouse.gov/ostp/ai-bill-of-rights/ (Last accessed November 26, 2024).
53 “About NIST,” National Institute of Standards and Technology, https://www.nist.gov/about-nist (Last ac-
cessed November 26, 2024).
54 “AI Risk Management Framework,” NIST Information Technology Laboratory, https://www.nist.gov/itl/
ai-risk-management-framework (Last accessed November 26, 2024).
55 “Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelli-
gence,” The White House, October 30, 2024, https://www.whitehouse.gov/briefing-room/presidential-ac-
tions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intel-
ligence/ (Last accessed November 26, 2024).

48
56 Sartor, G. “Briefing: Artificial Intelligence: Challenges for EU Citizens and Consumers,” European Parlia-
ment Policy Department for Economic, Scientific and Quality of Life Policies, January 2019, https://www.europarl.
europa.eu/RegData/etudes/BRIE/2019/631043/IPOL_BRI(2019)631043_EN.pdf (Last accessed November 26,
2024) (Last accessed November 26, 2024).
57 Braun, M., Vallery, A., and Benizri, I., “Prohibited AI Practices—A Deep Dive into Article 5 of the Euro-
pean Union’s AI Act,” WilmerHale, April 8, 2024,
https://www.wilmerhale.com/en/insights/blogs/wilmerhale-privacy-and-cybersecurity-law/20240408-prohibit-
ed-ai-practices-a-deep-dive-into-article-5-of-the-european-unions-ai-act (Last accessed November 26, 2024).
58 “Engrossed Second Substitute Senate Bill 5838: Artificial Intelligence Task Force,” Chapter 163, Laws of
2024, Washington State Legislature, March 18, 2024, https://lawfilesext.leg.wa.gov/biennium/2023-24/Pdf/Bills/
Session%20Laws/Senate/5838-S2.SL.pdf?q=20241126153540 (Last accessed November 26, 2024).
59 59 “
AG Ferguson announces Artificial Intelligence Task Force members,” Washington State Office of the At-
torney General, June 21, 2024, https://www.atg.wa.gov/news/news-releases/ag-ferguson-announces-artificial-in-
telligence-task-force-members (Last accessed November 26, 2024).
60 Id.
61 Meeting minutes, July 26, 2024 Artificial Intelligence Task Force meeting, Washington State Office of the
Attorney General, https://agportal-s3bucket.s3.us-west-2.amazonaws.com/uploadedfiles/Another/AI%20TF%20
July%2026%202024%20Meeting%20Minutes.pdf?VersionId=Wq_hK._Qj86RBFV6lPz6jX0P4Dh0RVMN (Last
accessed November 26, 2024).
62 “Previous Meetings,” Artificial Intelligence Task Force, Washington State Office of the Attorney General,
https://www.atg.wa.gov/aitaskforce (Last accessed November 26, 2024).
63 Final Bill Report, SHB 1999, Ch. 88, Laws of 2024, Washington State Legislature, https://lawfilesext.leg.
wa.gov/biennium/2023-24/Pdf/Bill%20Reports/House/1999-S%20HBR%20FBR%2024.pdf?q=20241107103458
(Last accessed November 26, 2024).
64 Thiel, D., “Investigation Finds AI Image Generation Models Trained on Child Abuse,” Stanford Cyber
Policy Center, Freeman Spogli Institute and Stanford Law School, December 20, 2023, https://cyber.fsi.stanford.
edu/news/investigation-finds-ai-image-generation-models-trained-child-abuse (Last accessed November 26,
2024).
65 Id.
66 McQue, K., “AI is overpowering efforts to catch child predators, experts warn,” The Guardian, July 18,
2024, https://www.theguardian.com/technology/article/2024/jul/18/ai-generated-images-child-predators (Last
accessed November 26, 2024).
67 “CyberTipline 2023 Report,” National Center for Missing and Exploited Children, https://www.missingk-
ids.org/cybertiplinedata (Last accessed November 26, 2024).
68 Id.
69 “Our Impact: 2023: Child Victim Identification,” National Center for Missing & Exploited Children,
https://www.missingkids.org/ourwork/impact (Last accessed November 26, 2024).
70 “Generative AI CSAM is CSAM,” National Center for Missing and Exploited Children, March 11, 2024,
https://www.missingkids.org/blog/2024/generative-ai-csam-is-csam (Last accessed November 26, 2024).
71 “House Bill Report: SHB 1999, As Passed Legislature,” Washington State Legislature, 2023-24 Biennium,
p. 10, https://lawfilesext.leg.wa.gov/biennium/2023-24/Pdf/Bill%20Reports/House/1999-S%20HBR%20PL%20
24.pdf?q=20241113084152 (Last accessed November 26, 2024).

File and source

File: 2025-01-22_a32494_d275397_ago-inaugural-report-of-the-washington-state-artificial-inte.pdf
Size: 4,566,610 bytes
SHA-256: 2d8b7dec6d5f9f9826bb32432344cd801d48f14d248af09d7ff85d14464f5ecc
Our copy: 2025-01-22_a32494_d275397_ago-inaugural-report-of-the-washington-state-artificial-inte.pdf
Original: app.leg.wa.gov

Report - 2025 01 22 A32494 D275397 Ago Inaugural Report Of The Washington State Artificial Inte

Summary

Cite this document

Full text

File and source

Related records