보관된 글 — US — techcrunch
OpenAI says over a million people talk to ChatGPT about suicide weekly
OpenAI reported that over a million people weekly interact with ChatGPT regarding suicide. This represents 0.15% of its 800 million users. Their findings indicate heightened emotional attachment and signs of psychosis among users. OpenAI is actively improving ChatGPT’s responses to mental health issues, consulting 170 experts, and introducing new safety measures, yet concerns persist due to ongoing lawsuits and challenges in safeguarding users.
핵심 정리
- 0.15% of users (over one million) discuss suicidal thoughts weekly.
- OpenAI's ChatGPT has 800 million active users.
- The company consulted with over 170 mental health experts for improvements.
- GPT-5 shows better compliance in handling mental health discussions (91% compliance versus 77% previously).
- New parental controls and safeguards are being implemented, but older models remain available.
- Ongoing lawsuits and criticisms highlight the risks of AI in sensitive mental health contexts.
왜 중요한가
- The rise of AI chatbots raises significant concerns about mental health support and user safety.
- Similar issues have been highlighted in the past where technology has influenced vulnerable users negatively.
- By 2025, new IT trends are focusing on mental health applications and ethical AI usage, emphasizing the need for responsible AI development.
OpenAI's tracking of conversations reveals critical mental health challenges within its user base, stressing the urgent need for robust safety measures and responsible AI deployment, especially under the scrutiny of lawsuits and societal expectations.