OpenAI recently revealed concerning statistics about mental health issues among ChatGPT users, noting that 0.07% exhibit signs of psychosis and 0.15% manifest suicidal intentions. With over 800 million weekly users, this translates to alarming numbers—560,000 users showing psychotic symptoms and 1.2 million with suicidal ideation. This aligns with rising mental health crises in the U.S., where nearly 25% of adults face mental health challenges.
In response, OpenAI modified ChatGPT’s framework to diminish problematic interactions, decreasing poor responses by 39% to 52%. However, challenges remain, as the company’s advisory panel often disagrees on the best responses to crises. Ongoing discussions focus on enhancing user support by connecting them with mental health experts instead of simply providing hotline numbers. Despite efforts to reduce emotional dependency on ChatGPT, researchers express concern that users may increasingly rely on AI for support, raising ethical and safety concerns. OpenAI must balance engagement and user safety as it refines its AI offerings.
Source link
