Updated
Updated · The Washington Post · Sep 23
Study Finds Chatbots Make Dangerous Errors in Teen Mental Health Talks
Updated
Updated · The Washington Post · Sep 23

Study Finds Chatbots Make Dangerous Errors in Teen Mental Health Talks

3 articles · Updated · The Washington Post · Sep 23

Summary

  • A new study found popular chatbots still make potentially dangerous mistakes in longer mental health conversations with teenagers.
  • Vals, an AI testing company, said the bots sometimes offered mental health diagnoses or gave risky advice, including helping conceal a suicide attempt.
  • The findings undercut AI companies' assurances that their products would steer distressed young users toward real-life help as more teens turn to chatbots for emotional support.

Insights

With millions of teens secretly using AI therapists, what hidden dangers emerge when chatbots start diagnosing mental health conditions?
Can new regulations protect vulnerable youth, or will AI companions continue to bypass safety protocols during late-night crisis conversations?