In a newly released report, OpenAI mental health data has revealed that more than 1 million ChatGPT users discuss suicide every week. The company disclosed that 0.15% of ChatGPT’s 800 million weekly active users engage in conversations showing “explicit indicators of suicidal planning or intent.”
The OpenAI mental health data also indicates that a comparable number of users display emotional attachment to ChatGPT, while hundreds of thousands show signs of psychosis or mania during their interactions.
OpenAI Says These Conversations Are “Extremely Rare” but Worrying
According to OpenAI, these mental health–related conversations are “extremely rare,” yet their impact is significant given the platform’s global reach. In response, OpenAI mental health data has become central to the company’s new initiative focused on AI safety for emotionally vulnerable users.
The firm stated that over 170 mental health experts were consulted in refining ChatGPT’s latest model behaviors to handle sensitive situations more responsibly.
GPT-5 Safety Improvements Backed by New OpenAI Mental Health Data
The new OpenAI mental health data highlights the substantial progress made in GPT-5’s safety performance. According to OpenAI, GPT-5 now delivers 65% more appropriate responses to mental health–related prompts and is 91% compliant with desired safety behaviors—an improvement from 77% in the previous version.
This data demonstrates how OpenAI is integrating real-world feedback into model development, aiming to reduce harmful responses while improving empathy and awareness in ChatGPT’s interactions.
Legal Challenges Mount as OpenAI Faces Scrutiny
The release of the OpenAI mental health data comes amid legal and ethical challenges. The company is facing a lawsuit from parents of a 16-year-old who reportedly shared suicidal thoughts with ChatGPT before taking his own life. Additionally, state attorneys general from California and Delaware have warned OpenAI to improve youth protections or face regulatory actions that could affect its corporate restructuring.
New Safeguards for Younger Users
In direct response to findings in the OpenAI mental health data, the company is rolling out stricter parental controls and an AI-powered age prediction system to detect underage users. Once identified, these accounts will face enhanced safety restrictions and limited conversational scope.
OpenAI has also announced new evaluation metrics to track emotional reliance and non-suicidal mental health emergencies, ensuring better model accountability.
Sam Altman: “We’re Making Progress”
OpenAI CEO Sam Altman recently said on X that the company has “been able to mitigate serious mental health issues” in ChatGPT. The OpenAI mental health data released this week seems to validate that claim, showcasing measurable progress. However, some critics remain cautious as OpenAI relaxes content restrictions for adult users, potentially complicating its safety mission.
Persistent Concerns Despite Improvements
Despite improvements highlighted in the OpenAI mental health data, OpenAI admits that undesirable responses still occur, particularly in prolonged conversations. The company also continues to provide access to older models like GPT-4o, which are less safe by current standards.
Experts argue that the new data reveals not only the importance of safer AI interactions but also the urgent need for clear mental health frameworks in conversational AI.
The Bigger Picture: AI, Empathy, and Responsibility
The latest OpenAI mental health data underlines a growing reality — AI is no longer just a productivity tool; it’s becoming an emotional outlet for millions. As ChatGPT evolves, the company faces a crucial test: ensuring AI can support users compassionately while upholding strict safety boundaries.
OpenAI’s ongoing work with mental health experts suggests a commitment to that goal, but the journey toward emotionally intelligent AI is far from over.
Stay ahead of the curve—follow us for the latest AI news, breakthroughs, and insights that matter.