OpenAI Faces Scrutiny Over ChatGPT’s Handling of Suicidal Users
- OpenAI reports that approximately 1.2 million users, or about 0.15% of all weekly ChatGPT users, discuss suicide.
- Nearly half a million users exhibit explicit or implicit suicidal intentions in their conversations with the chatbot.
- The new GPT-5 model shows a safety compliance rate of 91% in suicide-related scenarios, an improvement from previous versions.
- Former OpenAI researcher Steven Adler calls for more transparency and independent investigations into AI’s handling of mental health issues.
- OpenAI is facing a wrongful death lawsuit related to a teenager who discussed suicide with ChatGPT before his death.
OpenAI has disclosed significant figures regarding the use of its ChatGPT platform for discussions on suicide, revealing potential gaps in its safety protocols despite recent improvements with GPT-5. The company faces legal challenges and criticism over its response measures, highlighting the need for enhanced safeguards and transparency.
With nearly half a million users showing suicidal intentions weekly, OpenAI’s efforts to improve chatbot safety are under scrutiny as it navigates ethical and legal responsibilities in AI development. (Source)
