In brief
- OpenAI has delayed its adult content mode to early 2026, missing a December launch target set by CEO Sam Altman.
- The holdup centers on an AI age-prediction system meant to protect minors without misclassifying adults.
- The delay frustrates users as rivals already allow NSFW interactions, leaving OpenAI more cautious and constrained.
Looking forward to OpenAI’s erotic evolution? You’ll have to wait a bit longer.
OpenAI is pushing back the company’s adult content feature to at least early 2026, breaking a high-profile December deadline that CEO Sam Altman set just two months ago.
As first reported by The Verge, Fidji Simo, OpenAI’s CEO of applications, told reporters Thursday during the GPT-5.2 briefing that the company’s adult mode will now debut sometime in the first quarter of next year. The delay stems from ongoing work to perfect an age prediction system that OpenAI says needs to accurately identify teenagers without incorrectly flagging adults.
The announcement follows Altman’s October X post that ChatGPT would allow erotica for verified adults starting in December. That timeline was part of what Altman called OpenAI’s principle to “treat adult users like adults” after years of increasingly restrictive content policies designed to address mental health concerns.
The delay arrives at an awkward moment for OpenAI, which faces mounting pressure from frustrated users. Over 3,000 people signed a Change.org petition demanding an “Adult Mode” after ChatGPT’s filters began blocking even discussions about “kissing and non-sexual physical intimacy.”
Creative writers and romance novelists have complained for months that the chatbot shuts down legitimate storytelling attempts with policy violation warnings.
Meanwhile, OpenAI’s competitors haven’t waited. Elon Musk’s Grok chatbot already offers what xAI markets as “waifu companions”. The firm has nerfed its horniness, sure—and Grok users are angry—but it’s still a case of an AI behemoth taking NSFW content seriously.
Character.AI built a sizable user base around romantic AI interactions, despite facing lawsuits over teen safety. And open-source models like Llama run locally without corporate oversight, giving users complete freedom to generate whatever content they want, plus Meta allows its models to flirt with users—even teenagers.
The age verification system at the heart of the delay uses AI to estimate user age based on how people interact with ChatGPT. If the system can’t make a confident determination, then it defaults to the under-18 experience.
OpenAI’s caution reflects recent scrutiny over AI’s impact on vulnerable users. The company is already facing a wrongful death lawsuit from parents who claim ChatGPT contributed to their 16-year-old son’s suicide. Earlier this year, OpenAI formed an Expert Council on Well-Being and AI to advise on mental health impacts, and even the FTC launched an investigation into how tech companies protect minors.
For now, users eager to talk dirty to ChatGPT will need to wait until at least January—or use an open-source model and go crazy.
Generally Intelligent Newsletter
A weekly AI journey narrated by Gen, a generative AI model.
