Transparency Coalition counts 115 AI measures across 30 states, up nearly 58%

State AI legislation grew sharply in the Transparency Coalition’s new annual tally—but its headline count is not quite a count of enacted laws. Released October 27, the report tracks 115 measures across 30 states, including a small number awaiting governors’ decisions. Its findings put chatbot protections and human judgment in healthcare among lawmakers’ leading priorities.

A larger tally, with a defined cutoff

The 2026 State AI Legislation Report covers measures passed or enacted between January 1 and October 15. Its total is nearly 58% higher than the 73 AI-related laws included in the coalition’s 2025 report. That comparison describes the coalition’s tracked measures, rather than a uniform set of requirements adopted by every state.

The review examines AI development, design, deployment, transparency, accountability and use. Data center development measures are excluded. Inclusion also does not signal the coalition’s endorsement or opposition. The publisher itself advocates policies focused on AI safety, transparency and the public good, with particular attention to protecting children.

Chatbot behavior and healthcare decisions

Fourteen states advanced chatbot safety measures, according to the report. Those measures address self-harm and suicidal ideation, sexually explicit content, emotionally manipulative engagement and other potentially harmful interactions with minors. The coalition singles out California, Oregon and Washington as having enacted particularly extensive protections, including safeguards against tactics designed to keep children interacting with chatbots.

Other findings concern both the design of digital products and the decisions AI is allowed to make. Healthcare provisions distinguish between using automated tools and allowing those tools to act without human judgment. The report identifies several approaches:

• Insurance decisions: Alabama, Colorado, Georgia, Illinois, Indiana, Iowa, Utah and Washington enacted laws limiting insurers’ ability to rely solely on AI or automated systems for adverse medical authorization or coverage decisions.
• Mental health services: Colorado, Maine, Rhode Island, Tennessee and Vermont prohibited or restricted mental health therapy offered through AI chatbots, using states’ existing authority over licensed healthcare professions.
• Compulsive engagement: California, Colorado, Connecticut, Idaho, Illinois, Minnesota and South Carolina adopted measures addressing addictive design in products used by children and teens, including algorithmic feeds, autoplay and rewards.

Outside audits join disclosure requirements

For developers of the most advanced AI models, the report highlights Illinois’s AI Safety Measures Act. It requires risk-management frameworks and annual independent third-party audits. California enacted measures establishing standards and infrastructure for independent AI auditing organizations. The two findings concern different parts of oversight: requirements placed on developers and the organizations that assess AI systems.

Transparency rules also reach what people see. States required clearer disclosure of AI use in advertising, customer service, campaign communications and chatbot interactions. Washington enacted a requirement for covered AI developers to embed provenance information—information about a media item’s origin—in AI-generated or altered media. Separately, the report identifies strengthened protections against harmful or deceptive synthetic media in eight states.

In its forecast for 2027, the coalition expects lawmakers to revisit chatbot safeguards because the 2026 measures vary significantly. It also anticipates more attention to independent auditors’ certification, licensing and standards. Another potential route is public purchasing: the coalition expects agencies and schools’ procurement requirements to become a lever for stronger vendor safety, testing, transparency and accountability. Those are forecasts, not newly enacted obligations.