States ramped up efforts to regulate artificial intelligence in health care this year, moving from exploratory debates to concrete legislation.

Lawmakers in 46 states introduced more than 250 bills addressing AI’s role in health care. Seventeen states enacted 27 of those measures, with three more awaiting California Gov. Gavin Newsom’s (D) signature, according to consulting firm Manatt.

It is a sharp shift from previous years, when only a handful of state AI bills touched on health care — and most of those simply set up study commissions. Analysts expect the momentum to continue as lawmakers gain a clearer sense of the technology’s potential and risks.

In Pennsylvania on Monday, a bipartisan group of lawmakers introduced legislation that would provide new regulations for how AI is utilized and reported by insurers, hospitals and clinicians.

“As AI use has grown in the health care industry, we have already seen evidence that AI usage can reinforce bias and discrimination,” Pennsylvania Rep. Arvind Venkat (D), a physician, said in a statement. “This will allow us to ensure that insurers, clinicians and hospitals use AI effectively and do not use it to perpetuate potentially harmful biases in the medical field.”

Providers and health plans are affected the most by the legislation that regulates AI in health care, although some states have passed broader AI bills that also impact developers, Manatt partner Randi Seigel said during a recent webinar. This year’s measures largely fell into four categories, she said:

• Transparency: The largest number of bills — more than 90 — require that doctors and patients are told when they are interacting with AI systems. Some bills also require AI developers to obtain licenses from the state. At least 18 of those proposals contain language from a comprehensive AI law Colorado Gov. Jared Polis (D) signed in 2024, although one was vetoed and the rest have not been passed. 

• Anti-discrimination: More than 50 of the proposed bills include requirements to prohibit or address discrimination by AI tools. That includes a Maryland law, passed in May, that requires insurance carriers and health benefit plans to ensure that coverage decisions that rely on AI tools don’t result in discrimination. And a sweeping Texas law passed in May included a prohibition on the development or deployment of AI systems that discriminate.

• Payor use: Five states passed legislation effectively limiting insurers’ use of artificial intelligence to deny coverage for medical care. Those laws — in Arizona, Connecticut, Maryland, Nebraska and Texas — followed a first-in-the nation bill passed in California in 2024. They were among 60 bills in 25 states addressing concerns about the rise in automated insurance denials driven by AI-powered algorithms.

• Clinical contexts: About 20 proposed bills focused on how AI is used in clinical contexts, ranging from administrative tasks to direct communication with patients. One area of concern is the expanding market for mental health chatbots. Utah Gov. Spencer Cox (R)  in March signed the nation’s first law aiming to regulate mental health chatbots. Bills were also introduced in California, Illinois, Nevada, New Jersey, New York, North Carolina and Texas.

Analysts, lawmakers and health care interest groups predict even more activity at the state level, as federal regulation efforts have stalled and the technology becomes more ingrained in the health care system. 

A Florida House Commerce subcommittee scheduled a hearing Tuesday to examine the extent of AI’s use in the insurance industry.

The Pennsylvania bill, which was referred to the House Communications and Technology Committee, would require insurers, hospitals and clinicians to tell patients and the public when they are using AI; ensure that a human signs off on any AI-aided decision made by insurers, hospitals or clinicians; and require that insurers, hospitals and clinicians demonstrate to state regulators that they have taken steps to minimize illegal bias and discrimination in their use of AI. 

“With the introduction of AI, we need experienced doctors and nurses even more now to assess the accuracy of AI to ensure that bias and discrimination haven’t influenced its findings,” said Pennsylvania Rep. Bridget Kosierowski (D), a nurse and cosponsor of the bill.

Physicians’ use of AI for certain tasks nearly doubled in the last year, according to a survey released this summer by the American Medical Association, which has called for more oversight of the industry.  

“For now, I think a little bit about the analogy about the rise of the internet,” Oregon Sen. Lisa Reynolds (D) said at a recent roundtable discussion with lawmakers and doctors. “We certainly have harnessed that for good. But I would argue that there are some places we really missed the boat, especially when it comes to youth mental health and the role of social media companies.”

Reynolds said AI could help tackle some of the most pressing challenges in health care — from lowering costs and expanding access to supporting the workforce and driving economic growth — if policymakers strike the right balance between guardrails and innovation.

Oregon Rep. Daniel Nguyen (D), who also joined the roundtable, said the legislature is ready to move forward after convening a 2024 AI task force. With its tech talent and medical research industry, the state could become a national leader in health AI, he said. 

“There’s just so much opportunity that can be achieved if we think creatively,” Nguyen said.