With Congress still yet to pass a federal AI law, US states have become the main engine of AI regulation. A report released on October 27 by the nonprofit Transparency Coalition counts 115 AI-related laws passed or enacted across 30 states between January 1 and October 15, 2026. That is nearly 58% more than the 73 laws the group tracked in its 2025 report, and the clearest picture yet of how fast state legislatures are writing the rules for chatbots, insurers, schools and frontier model developers.
Chatbots and Children Lead the Agenda
Protecting minors from harmful AI companion chatbots was the top legislative priority of the year. The report finds that 14 states advanced measures dealing with self-harm content, sexual content, emotional manipulation and engagement tactics. California, Oregon and Washington enacted the most extensive protections.
A related wave targeted addictive design. California, Colorado, Connecticut, Idaho, Illinois, Minnesota and South Carolina passed measures addressing compulsive features for minors, including algorithmic feeds, autoplay and reward mechanics. The overlap shows lawmakers increasingly treating chatbots and social feeds as part of the same child-safety problem.
Keeping Humans in Healthcare Decisions
The second major theme was preserving human judgment in healthcare. Eight states, Alabama, Colorado, Georgia, Illinois, Indiana, Iowa, Utah and Washington, limited health insurers' ability to rely on AI alone when denying coverage or prior authorisation.
Five states, Colorado, Maine, Rhode Island, Tennessee and Vermont, restricted AI chatbots from offering mental health therapy, using existing professional licensing authority rather than creating new AI-specific regulators. That approach could become a model for other licensed professions.
Deepfakes, Disclosure, Schools and Frontier Models
Beyond the headline issues, the report documents broad activity across several other areas:
- Deepfakes: California, Hawaii, Kansas, Louisiana, Maryland, Tennessee, Utah and Washington strengthened protections against harmful or deceptive synthetic media
- Disclosure: States required clearer AI disclosure in advertising, customer service, campaign communications and chatbot interactions; Washington now requires covered developers to embed provenance information in AI-generated or altered media
- Education: California, Illinois, Maryland, Oklahoma, Tennessee, Utah and Virginia addressed AI procurement, classroom use, device policies and the role of human teachers
- Frontier models: Illinois enacted the AI Safety Measures Act, requiring risk-management frameworks and annual independent third-party audits, while California set standards for independent AI auditing organisations
The report excludes data centre measures, and the Coalition notes that inclusion of a law does not signal support or opposition. A handful of measures were still awaiting governors' signatures because of late signing deadlines.
"State lawmakers in both parties have stepped up to the challenge of regulating AI for the good of our communities," said Rob Eleveld, the Coalition's chief executive. He added that the seriousness of this year's efforts stood out.
What 2027 Could Bring
The Coalition expects the next legislative season to focus on:
- AI product liability, defining who is responsible when AI systems cause harm
- Second-generation chatbot safeguards that revisit the first wave of laws
- Certification, licensing and standards for independent AI auditors
- Procurement standards for AI vendors selling to state agencies and school systems
- Limits on AI in other state-licensed professions
- Testing, deployment and monitoring requirements for advanced AI models
The pace of change is notable on its own. A 58% jump in a single year means many organisations that reviewed their AI obligations in 2025 are already working from an outdated map. Legal and compliance teams will need to track not only new statutes but also how states that legislated early, such as California and Colorado, revise their first-generation rules as practical problems emerge.
Why It Matters
The report underlines a growing tension in US AI policy. While Washington debates federal legislation, including proposals that would pre-empt some state rules, the states have already built a substantial and expanding body of law. Companies deploying chatbots, AI-assisted claims decisions or generative media in the US now face a patchwork of obligations that differ by state, from provenance labelling in Washington to third-party audits in Illinois.
For AI developers and enterprises, the practical takeaway is that compliance can no longer wait for a single national standard. Child safety, healthcare decisions and synthetic media are where enforcement risk is highest today, and auditing and liability rules are likely next. Bipartisan momentum at the state level also suggests these rules are not a passing trend: lawmakers in both parties are treating AI harms as a mainstream consumer protection issue.
