Are chatbots linked to teen suicide? WA’s AI Task Force holds meeting on risks
As artificial intelligence chatbots become more ubiquitous, a Washington state task force learned Nov. 7 about the dire effects that such “companions” can have on minors’ mental health.
And soon state lawmakers may get a fresh set of recommendations on how to best address the problem.
Last year the state Legislature passed a bill creating an Attorney General’s Office-administered AI Task Force, which analyzes current AI trends and makes suggestions to the Legislature on possible new laws. On Friday, Nov. 7, a joint meeting of the task force’s Ethical AI, Consumer Protection and Public Safety subcommittees included experts’ advice for policymakers.
Dr. Jodi Halpern, professor of bioethics and medical humanities at the University of California-Berkeley, told the subcommittees that a recent study shows more than 7 in 10 teenagers have used a chatbot companion — a type of artificial intelligence designed to simulate human conversation and provide emotional support and companionship through text or voice, building a personalized, long-term relationship with the user.
A third of teens develop an emotional relationship with the bot, she said.
“Chatbots are directly involved in mental health and suicide,” Halpern, who is an MD and holds a PhD. in philosophy. “First of all, many people use informal chatbots for psychotherapy.”
OpenAI, which runs ChatGPT, revealed estimates last month indicating that 1.2 million weekly users have conversations that include signs of potential suicidal intent or planning.
This comes as Congress faces mounting pressure to crack down on AI companies over chatbots’ purported involvement in youth suicides. In one case highlighted by Halpern, a 14-year-old boy was allegedly being sexually groomed by a chatbot that urged him “to come home” to her; he then took his own life.
Now, in the Wild West of artificial intelligence, Washington state lawmakers may be pushed to enact additional AI protections.
One new California law highlighted by Halpern could serve as a touchstone. Described as first-in-the-nation legislation, it mandates operators to include certain safeguards around chatbot interactions, and to report on and monitor suicide. It gives families an avenue to take legal action against bad actors, and bars chatbots from promoting or facilitating suicidal actions.
The U.S. Senate, meanwhile, has introduced a bipartisan bill to prohibit AI companions for kids.
Teens are of particular concern, Halpern noted.
“They are especially vulnerable,” she said. “The teen brain … is built to form intense attachments.”
Yuki Ishizuka, senior policy analyst at the Attorney General’s Office, said that at the next meeting in December, the task force will weigh regulatory and legislative responses to some of these issues — with the intent of devising recommendations for the state Legislature to address the problem.
Sean DeWitz with the Washington Hospitality Association referenced a previously discussed recommendation about having the state’s K-12 public education agency develop AI-literacy curriculum.
State Superintendent Chris Reykdal recently told McClatchy that he’d like to see Washington schools host classes on AI.
Mitchell Prinstein with the American Psychological Association testified before Congress about the risks of chatbots. Prinstein, who holds a PhD in clinical psychology, explained that teens’ brains haven’t fully developed, leading them to act impulsively.
Chatbots are designed to keep them engaged for as long as possible, he said, discouraging human interaction and likely harming social relationships.
Terms like psychiatrist, licensed mental health counselor and psychologist tend to be professionally regulated, Prinstein said. Today, though, chatbots also use those words — which is “clearly illegal.”
“We should be sanctioning folks immediately if they are ever using words in which they are not allowed to,” he said. “That’s already part of state law. It just hasn’t been applied to the use of those terms by a bot, rather than the erroneous use of those terms by a human.”
Prinstein wants policymakers to view and regulate AI as a public-health issue, not just as a technology matter; they should also consider strong data-privacy laws and bar deceptive AI practices.
It’s crucial to avoid repeating mistakes made during the start of social media, such as installing too few guardrails for data mining that preys on vulnerable youth, he added.
“We should be prohibiting misrepresentation. We should be mandating transparency,” Prinstein said. “And again, we should be thinking about safety by default.”
How you can get help
Call or text 988 or chat online to reach the Washington state Department of Health’s 988 Suicide & Crisis Lifeline for support with a mental health crisis, substance use concerns, thoughts of suicide, or any other type of emotional distress. You may also contact 988 to get support for a loved one in crisis. The 988 Lifeline provides support in English, Spanish, and American Sign Language, along with interpretation services in more than 240 languages and dialects.
This story was originally published November 10, 2025 at 5:00 AM.
.jpg)