5838-S2.E AMH APP H3451.1
E2SSB 5838 - H COMM AMD
By Committee on Appropriations
ADOPTED 02/29/2024
Strike everything after the enacting clause and insert the following:
"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 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, 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.
(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."
Correct the title.
EFFECT: (1) Modifies the membership of the task force to also require: (a) One member representing private technology industry groups and one member representing business associations; (b) one member representing the LGBTQ+ community; (c) one member from the retail industry; (d) one member from the hospitality industry; and (e) three members, instead of one member, representing community advocate organizations that represent communities disproportionately vulnerable to being harmed by algorithmic bias.
(2) Expands the findings and recommendations the task force must make, including findings and recommendations related to: Ethical artificial intelligence; algorithmic discrimination issues; minimizing biased outputs or applications; prioritizing transparency; data privacy, security, collection, and retention practices; existing remedies for addressing identified potential harms and new means of enforcement and remedies if needed; and establishing an ongoing committee.
--- END ---