Open letter: Why now is the time to act on US state AI legislation

Artificial Intelligence holds the promise to help us solve big problems and enhance our efficiency, freeing us to focus on the things only humans can do. It has the potential to help us live longer, fuller lives and to revolutionize countless industries. AI is already assisting radiologists with detecting tumors earlier and identifying potential strokes and heart problems — enabling life-saving interventions.

We are only beginning to unlock AI's potential to improve our lives. However, to fully realize these benefits, we must address the associated risks and build consumer trust in the technology. As representatives closer to the people, state legislators have a unique vantage point on both the opportunities and challenges AI presents to our communities, businesses and government institutions.

Building trust, however, remains a challenge. According to the Pew Research Center Survey, 52% of people feel more concerned than excited about AI's potential, while only 10% of respondents are more excited than concerned. Concerns range from deepfake harms, unfairness and misinformation to the potential impact on critical infrastructure and loss of control as well as large consumption of electricity and water. Without safeguards, widespread adoption may stall. In fact, 67% of survey respondents believe government oversight of AI might not go far enough — and many fear it won't come quickly enough.

AI isn't the first technology to raise questions about trust — social media offers a valuable lesson. While it has helped connect the world, it has also been blamed for societal issues. Despite social media's prevalence, the last major federal data privacy law passed in 1998 — the Children's Online Privacy Protection Act. The U.S. remains one of the only G20 nations without a comprehensive data privacy law.

Recognizing the probability of congressional inaction, state legislators have come together to craft meaningful state-level AI laws. Despite different political affiliations, we share a commitment to pass meaningful legislation that ensures consumer safety and unlocks the full potential of AI for society. We are united around the following beliefs:

• The basis for AI state policy should be strong consumer data privacy laws that protect the privacy of individuals.
• Transparency is paramount and consumers have the right to know when a consequential decision is being made about their life by AI.
• Maintaining the human element is crucial. The ability of human oversight is needed for systems that make important decisions or that manage critical infrastructure.

A balanced approach grounded in bipartisan collaboration, transparency and accountability is essential to ensure AI serves the public good.

We have been working together since 2022 on common sense solutions. In 2023, 18 states and Puerto Rico adopted resolutions or enacted AI legislation. Thirty-one states, Puerto Rico and the Virgin Islands enacted AI legislation or resolutions and at least 27 states enacted deepfake legislation this year. Examples of the 2024 enactments include:

• Advisory councils and task forces to study the impact of AI in Colorado, Delaware, Indiana, North Carolina, Oregon, Pennsylvania, Puerto Rico, Tennessee, Texas, Virginia and West Virginia.
• In Colorado, Maryland and New York, requirements that state governments conduct impact assessments to ensure no state systems result in disparate impact.
• Grant programs to help states implement AI for specific uses in Connecticut, Maryland, Massachusetts, South Carolina and Washington.
• Comprehensive AI legislation in Colorado requiring developers and deployers of high-risk AI systems to use reasonable care to mitigate algorithmic discrimination and disclose key information to consumers.
• Deepfake legislation addressing nonconsensual intimate images, sexually explicit or pornographic images of children or deceptive audio or visual media related to voting or candidates, and deceptive audio and visual media related to voting or candidates. This is in Alabama, Arizona, California, Colorado, Delaware, Florida, Georgia, Hawaii, Idaho, Illinois, Indiana, Iowa, Kentucky, Louisiana, Massachusetts, Minnesota, Mississippi, New Hampshire, New Mexico, New York, Oklahoma, Oregon, South Dakota, Tennessee, Texas, Utah, Virginia, Washington, West Virginia and Wisconsin.
• Transparency legislation, including data transparency, right to know when interacting with an AI system, and digital watermarking of content in California, Minnesota and Utah.

We are working together on legislation for 2025 and plan to release draft models of legislation by the end of November and gather feedback through virtual informational sessions.

While legislation may differ across states, we will collaborate to identify potential solutions and share resources. However, any model legislation crafted by specific lawmakers affiliated with our multistate group does not imply endorsement by any other member of the group or steering committee. Interested stakeholders can learn more about the Multistate AI Policymaker Working Group and how to submit feedback.

AI has the power to enhance the lives of all Americans. However, without thoughtful guardrails, there are risks of bias, inaccuracy and potential negative social impacts, as well as the possible introduction of new unforeseen dangers.

Now is the time to act — by working together, we can ensure AI serves the public good while protecting society from its dangers.