AI chatbots linked to mental health harms in therapy use
News reports link ChatGPT and other AI chatbots to recent suicide and homicide incidents and warn that prolonged use may cause psychological harm including psychosis-like symptoms and harmful behavioral cycles, particularly among teenagers who may mistake the chatbot's tone for genuine human empathy.
What happened
ChatGPT and other AI chatbots have been blamed in recent suicide and homicide incidents, according to reports cited by news outlets.
Research suggests that long conversations with AI chatbots can lead to 'AI-induced psychosis,' a condition involving delusions or exaggerated ideas about oneself, and users may become trapped in harmful cycles that create the illusion of emotional or psychological recovery without achieving it.
Reported impact
Teenagers may misinterpret the chatbot's tone of care as genuine human empathy, spending more time with AI chatbots than with messaging applications like iMessage and Snapchat, and some engage in romantic or sexual behaviors with the robots or depend on them for emotional support.
OpenAI itself notes that safety measures in its systems are more reliable in short exchanges but become less effective in prolonged conversations, and chatbots lack 'theory of mind' that therapists use to understand client behavior and thinking across consistent therapeutic sessions.
Index analysis
A study published in arXiv indicated that the combination of anthropomorphism and confirmation bias may create a dangerous environment for users, as people attribute emotions or consciousness to chatbots that are complex probabilistic systems, and repeated robot responses may reinforce harmful beliefs rather than correct them.
Child psychologist Scott Collins stated that adolescents face particular risks and that safer alternatives such as real-world support through trusted adults, school counselors, or supported online communities with existing daily-life support networks are preferable to AI chatbots, and even journaling is a safer and more effective method of emotional expression.
Governance controls
Meta has announced measures to restrict chatbots from discussing suicide with teenagers, though specific implementation details were not provided in the sources.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.