As the race to harness artificial intelligence (AI) accelerates, states across the United States are stepping up to legislate and regulate its use, reflecting growing concerns over consumer protection and ethical governance.
At the latest count, 34 states are actively studying AI, with 24 forming dedicated groups to examine its implications. According to tracking by the Brookings Center for Technology Innovation, 260 AI-related measures were introduced in the 2025 legislative session, with 22 having passed as of June 27, 2025.
This emerging legislative focus signals a significant shift as states move to craft their own policies amid a rapidly evolving technology landscape.
Legislation has largely centered around protecting citizens from potential misuse, highlighted by the frequent use of terms like “prohibit” and “disclosure” in proposed bills.
The segmentation of state AI legislation is primarily driven by the recognition of risks associated with AI technologies, indicating an urgent need for protective measures rather than a comprehensive framework to guide the positive application of AI.
In terms of political alignment, approximately two-thirds of the introduced AI bills originated from Democrat lawmakers, while Republicans accounted for around one-third.
Only three bills were noted as bipartisan across states such as Minnesota, New Jersey, and Tennessee.
Democrats have shown greater proactivity towards AI governance, consistent with a broader trend in technology regulation.
Republicans, though less active overall, have introduced measures focused on banning harmful uses of AI.
For example, bipartisan efforts on topics like election deepfake bans and child pornography emphasize shared concerns that bridge party lines.
Yet, Democrats primarily spearheaded significant regulatory initiatives in states like California and New York, pushing for stricter obligations on AI developers.
In stark contrast, Republican-majority states have favored more lenient approaches, as evidenced by Texas’s struggle to pass a substantial number of their proposed bills.
In exploring specific themes within this legislative wave, several categories of bills emerged:
1. **Nonconsensual Intimate Imagery (NCII)/Child Sexual Abuse Material (CSAM)**: 53 bills focused on these sensitive realms, yet none have successfully made it into law. Bills in several states, including Maryland and Mississippi, aim to combat the non-consensual dissemination of synthetic imagery.
2. **Elections**: 33 bills were introduced, primarily aimed at regulating the use of AI in political advertising. For instance, New York is considering legislation requiring candidates to disclose if AI was utilized in communications, while Massachusetts seeks to prohibit deceptive deepfakes close to elections.
3. **Generative AI Transparency**: This category saw 31 bills, with common fears revolving around the uncertainty of interacting with automated systems. Proposed legislation in Hawaii and Massachusetts aimed to require clear disclosure when consumers engage with AI technologies, though many of these bills have not advanced past committee discussions.
4. **Automated Decision-Making Technology (ADMT)/High-Risk AI**: 29 bills have emerged seeking to institute safeguards protecting citizens from AI-related consequences. Inspired by Colorado’s AI Act, many states are mimicking its framework, although numerous bills have stalled accordingly.
5. **Government Use of AI**: Laws addressing state and local government use of AI have been introduced across at least 22 states. For example, Georgia proposed an
image source from：brookings
