Illinois becomes third state to restrict use of AI in mental health industry as experts warn about ‘AI psychosis’

Illinois passed a bill banning therapists from employing artificial intelligence chatbots for assistance with mental health therapy, as experts countrywide warn against people’s ever-growing reliance on the machines.

The “Therapy Resources Oversight” legislation prohibits licensed mental health professionals in Illinois from using AI for treatment decisions or communication with clients. It also bans companies from recommending chatbot therapy tools as a be-all alternative to traditional therapy.

Enforcement of the bill will rely on complaints from the public that the Illinois Department of Financial and Professional Regulation will investigate. Anyone determined to be violating the ban could face a civil penalty of up to $10,000, according to the legislation text.

Utah and Nevada, two Republican-run states, previously passed similar laws limiting AI’s capacity in mental health services in May and late June, respectively.

Unregulated chatbots can take harmless conversations in any direction, sometimes incidentally leading people into divulging sensitive information or pushing people who are already in vulnerable situations to do something drastic, like take their own life, experts have warned.

A Stanford University study released in June found that many chatbots, which are programmed to respond enthusiastically to users, fail to sidestep concerning prompts — including requests for high bridges in specific locations to jump off of.

Whereas chatbots affirm unequivocally regardless of the circumstance, therapists provide support and the means to help their patients improve, Vaile Wright, senior director for the office of health care innovation at the American Psychological Association, told the Washington Post.

“Therapists are validating, but it’s also our job to point out when somebody is engaging in unhealthy thoughts, feelings, behaviors and then help somebody challenge those and find better options,” Wright told the outlet.

The bans, though, are difficult to effectively enforce — and can’t prevent everyday people from turning to AI for mental health assistance on their own.

New research released in early August found that many bots like ChatGPT are inducing “AI psychosis” in unwitting users with no history of mental illnesses.

Roughly 75% of Americans have used some form of AI in the last six months, with 33% reporting daily usage for anything from help on homework to desperate romantic connections. This deep engagement is breeding psychological distress in heavy users, according to the digital marketing study.

Many youth, in particular, are falling down the chatbot rabbit hole and turning to machines to supplement human interaction.

Character.Ai, a popular platform where users can create and share chatbots usually based on fictional characters, had to place a warning clarifying that anything the bots say “should not be relied upon as fact or advice” after a Florida teen fell in love with his “Game of Thrones” AI character and took his own life.

The platform is still dealing with a lawsuit filed against the company for the teen’s death. Despite repeated attempts to dismiss it on First Amendment grounds, a federal judge ruled that the suit could move forward in August.

Another Texas family sued Character.Ai after a chatbot on the app named “Shonie” encouraged their autistic son to cut himself.