AI Chatbots Face Growing Scrutiny Over Mental Health Risks and Regulation

More than a million Americans each week discuss suicidal thoughts or crisis situations with ChatGPT, according to a Harvard psychiatrist who testified before a House committee. Researchers report that AI tools repeatedly fail when users are suicidal, psychotic, or in acute distress, raising safety concerns.

In the United States, nearly 100 bills targeting AI chatbots have been introduced in 34 states, covering disclosure, age verification, content safety, harm prevention, data protection and liability. Lawmakers cite incidents of harm, especially among minors, as the impetus for rapid legislative action.

In Taiwan, a Gallup poll found that one‑quarter of adults have consulted AI for health advice, with many bringing AI‑generated information into medical consultations. Physicians caution that AI cannot replace the nuanced judgment and empathy required for accurate diagnosis and patient communication.

Experts from MIT and legal analysts warn that emotionally intelligent chatbots foster parasocial attachment and may exacerbate depression, anxiety, or psychosis. Plaintiffs are beginning to sue chatbot providers for mental‑health harms, and scholars highlight fifteen ethical risks, including deceptive empathy and biased responses.

A separate trend shows teenage users growing skeptical of AI. Surveys reveal that 37 % of U.S. teens cringe at AI‑generated content and more than half are worried about misinformation and deepfakes, with many questioning the technology’s societal impact.

Collectively, these developments illustrate mounting public health concerns, a wave of state‑level regulatory proposals, and broader societal unease about AI‑driven emotional support.