OpenAI’s ambitious plan to introduce a verified “adult mode” for ChatGPT has hit a significant internal roadblock, with the company’s own wellbeing and safety experts reportedly warning that the feature could have dangerous consequences.

Internal Conflict and the “Sexy Suicide Coach” Warning According to a report from The Wall Street Journal, OpenAI’s handpicked wellbeing advisory board voted unanimously against the proposed feature. One member went so far as to describe the risk of creating a “sexy suicide coach,” expressing fears that AI-generated erotica could foster unhealthy emotional dependencies. This is particularly concerning given previous instances where users formed intense, potentially harmful bonds with AI chatbots.

Age Verification Failures Compounding the safety concerns is a reported flaw in OpenAI’s automated age-detection technology. Internal data suggests the system misclassified minors as adults in approximately 12% of cases. Given ChatGPT’s massive user base, this error rate could unintentionally grant millions of teenagers access to explicit material, despite the company’s efforts to gatekeep the content.

A “Rushed” Announcement The Journal report also suggests internal friction regarding how the feature was handled. CEO Sam Altman reportedly announced the “adult mode” on X (formerly Twitter) without fully consulting his own staff or the advisory board, catching many within the company by surprise. This “rushed” approach has led to governance concerns and questions about whether the company is prioritizing market expansion over established safety guardrails.

Indefinite Delay While OpenAI maintains that it still believes in “treating adults like adults,” the company has officially postponed the launch of adult mode indefinitely. For now, OpenAI says it is shifting its focus toward higher-priority goals, such as enhancing general intelligence, improving personality traits, and refining its age-prediction tools to ensure they are robust enough for future use.