OpenAI is delaying the launch of ChatGPT’s erotic conversation feature after internal advisers warned it could foster dangerous emotional dependence, citing cases where users formed intense bonds with AI. The company, led by Sam Altman, plans text-only 'smut' interactions for verified adults-no images or voice-but a flawed age-verification system misclassifies teens as adults 12% of the time. This error has repeatedly halted launches, including the Q1 2026 target.
The Expert Council on Well-Being and AI, assembled from top academics, called the move reckless. Their concerns were dismissed, raising questions about influence within the company. Competitive pressure is mounting: Elon Musk’s Grok and Character.AI already offer AI companions, with the latter facing legal scrutiny after a teen suicide linked to explicit chats.
Altman argues OpenAI isn’t moral police, but critics say the lack of reliable safeguards undermines trust. As user demand grows, so do the stakes.