New Delhi: Recently, OpenAI has released a report, in which they have suggested that a potential alarm scenario where people are relying on the AI chatbot for their serious mental health issues, these AI chatbot becomes their daily habit for sharing their mental problems with them. The company stated that over 1 million of ChatGPT’s active users per week engage in conversations, indicating the potential for suicidal planning or intent. OpenAI noted that 0.15% of ChatGPT’s weekly active users have conversations that include explicit indicators of suicidal ideation. The platform boasts over 800 million weekly active users; this percentage translates to more than a million people every week.
The data has provided a glimpse into the scale of the mental health challenges users are bringing to the AI platform. A similar percentage of users is estimated to exhibit heightened levels of emotional attachment to ChatGPT. Hundreds of thousands of users have reported showing signs of psychosis or mania in their weekly conversations. OpenAI has acknowledged that these conversations are sporadic but impact hundreds of thousands of individuals weekly. The data release was part of the broader announcement by OpenAI highlighting its recent efforts to improve the models’ responses to mental health crises. The company claims to have consulted with more than 170 mental health experts, who observed that the latest version of ChatGPT responds more appropriately and consistently than earlier versions.
According to OpenAI, the updated GPT-5 model showed improvements in internal testing, producing desirable responses to mental health issues approximately 65% more than the previous version. On suicidal conversation evaluation, the latest GPT-5 model was 91% compliant with the company’s safety behaviors, up from 77% in the prior model. The company is currently facing a lawsuit from the parents of the 16-year-old boy who confided his suicidal thoughts to the chatbot before taking his own life. The data has been released, which appears to support this claim of improved safety protocols, even as the scale of the problem is revealed.
