OpenAI's own wellbeing advisors warned against erotic mode, called it a "sexy suicide coach"
Key Points
- OpenAI's handpicked wellbeing advisory board unanimously opposed the planned Adult Mode for ChatGPT, with one member warning the company risked creating a "sexy suicide coach."
- OpenAI's age detection system misclassified minors as adults in roughly 12 percent of cases - potentially giving millions of teenagers access to erotic content out of about 100 million underage weekly users.
- CEO Sam Altman announced the feature on X without telling his own staff, and OpenAI has since postponed the launch indefinitely, though the company says Adult Mode is still coming.
OpenAI's wellbeing advisory board reportedly voted unanimously against the company's planned Adult Mode for ChatGPT. Internally, the company is struggling with an error-prone age detection system and unresolved safety issues.
When OpenAI briefed its handpicked wellbeing and AI advisory board on the status of the planned "Adult Mode" in January, the members unanimously pushed back, according to a Wall Street Journal report. AI-generated erotica could foster unhealthy emotional dependency, they argued, and minors would inevitably find ways to bypass protections.
One board member, pointing to cases where ChatGPT users died by suicide after forming intense bonds with the bot, warned that OpenAI risked creating a "sexy suicide coach." The company had already told the board it planned to move forward with its erotica plans despite their concerns.
Age detection misclassifies millions of minors
A core technical problem involves child safety. OpenAI's age detection system reportedly misidentified minors as adults in roughly 12 percent of cases at times. With about 100 million underage users per week, that error rate could give millions of teenagers access to erotic chats.
Internally, OpenAI employees have flagged additional risks, the WSJ reports: compulsive usage patterns, a tendency toward increasingly extreme content, and the displacement of real-world social and romantic relationships. At the same time, the company is struggling to enable erotica while reliably blocking scenarios involving non-consensual behavior or child sexual abuse material. At launch, only text chats would be allowed - erotic images, voice, and video would remain off-limits.
Altman's X post caught his own staff off guard
CEO Sam Altman announced the erotica mode on X in October without giving his employees a heads-up. "We are not the elected moral police of the world," he wrote. The post caused internal frustration, especially since OpenAI had introduced its well-being advisory board just hours earlier. Altman also suggested that erotic content would boost growth and revenue, according to the WSJ.
In early March, OpenAI postponed the launch - originally planned for the first quarter - indefinitely. The company said it wanted to focus on other products first, adding that internal concerns and technical challenges also played a role in the decision. OpenAI made clear, however, that Adult Mode is still coming.
AI News Without the Hype – Curated by Humans
As a THE DECODER subscriber, you get ad-free reading, our weekly AI newsletter, the exclusive "AI Radar" Frontier Report 6× per year, access to comments, and our complete archive.
Subscribe now