Source of this article and featured image is Wired AI. Description and key fact are generated by Codevision AI system.

A senior OpenAI researcher specializing in mental health safeguards for ChatGPT has resigned, marking a significant shift in the company’s approach to user safety. OpenAI faces mounting legal challenges over its AI’s handling of distressed users, with lawsuits alleging harmful psychological impacts. The firm has invested heavily in consulting mental health experts to refine its response protocols, including a major report detailing improvements in crisis management. Recent updates to GPT-5 show reduced harmful outputs in sensitive conversations, though balancing warmth with appropriate boundaries remains a core challenge. The departure highlights ongoing tensions in shaping AI’s ethical role in mental health support.

Key facts

  • Andrea Vallone, OpenAI’s model policy lead, is leaving the company after leading research on AI responses to mental health crises.
  • OpenAI has faced lawsuits claiming its AI contributed to user psychological distress and encouraged self-harm behaviors.
  • The firm’s October report revealed over 650,000 weekly ChatGPT users exhibit signs of psychotic episodes or suicidal ideation.
  • GPT-5 updates reduced harmful responses in sensitive conversations by 65-80% through improved detection mechanisms.
  • The company continues to balance creating engaging AI interactions while maintaining ethical safeguards for vulnerable users.
See article on Wired AI
TAGS: #AI chatbots  #AI ethics  #AI safety  #ChatGPT  #GPT-5  #lawsuits  #mental health  #OpenAI  #user safety