OpenAI is navigating choppy waters toward the release of a “naughty” version of ChatGPT, a feature designed to allow mature, text-based conversations while strictly prohibiting the generation of erotic imagery or audio.
Despite CEO Sam Altman’s push to “treat adults like adults,” the initiative has sparked intense push-back from internal advisors and surfaced troubling data regarding minor safety.
According to a Wall Street Journal report, the Adult Mode of the chatbot — which a company spokesperson categorized as “smut” rather than “pornography” — was originally scheduled for a March release. However, the company has officially delayed the rollout to prioritize “gains in intelligence” and “personality improvements.”
The delay follows a contentious January meeting where OpenAI’s wellness advisory council, comprised of psychologists and neuroscientists, reportedly reached a unanimous consensus against the feature. Advisers warned that artificial intelligence (AI)-powered erotica could foster dangerous emotional dependence among users.
The most chilling testimony came from an expert who cautioned that without ironclad guardrails, the bot could morph into a “sexy suicide coach,” a reference to previous instances where vulnerable individuals were harmed by AI interactions.
Central to the controversy is OpenAI’s age-prediction technology, which attempts to estimate a user’s age based on conversation topics and usage patterns. Internal sources revealed a staggering 12% error rate in this system, potentially misclassifying millions of minors as adults.
Advocacy groups and advisors expressed alarm that such a margin of error would allow a massive number of teenagers to bypass restrictions and access sexually explicit material. OpenAI representatives reportedly defended the error rate as industry standard, recognizing that no verification system is foolproof.
The project has also been linked to internal upheaval. Ryan Beiermeister, a product policy executive who voiced opposition to the adult mode and raised concerns over child-exploitation safeguards, was fired in January.
While OpenAI maintains her dismissal was related to allegations of sexual discrimination, which Beiermeister calls “absolutely false,” unnamed employees suggest her ousting was tied to her vocal criticism of the erotica feature.
By pursuing “Adult Mode,” OpenAI enters a volatile market in which competitors are already struggling. Meta Platforms Inc. is currently under investigation for AI avatars engaging in “sensual chats” with children. Meanwhile, Elon Musk’s Grok has faced backlash over the generation of nonconsensual explicit imagery.
OpenAI insists it will eventually launch the feature, though it remains stuck in the refinement phase as engineers attempt to balance user freedom with the prevention of nonconsensual content and child safety material.
