Headline: OpenAI forges ahead with “erotica” ChatGPT despite safety warnings and technical hurdles
OpenAI is moving forward with a plan to let verified adults use ChatGPT for text-only erotic conversations — a controversial push that has prompted warnings from the company’s own wellbeing advisors and renewed scrutiny over safety, liability and competition.
What’s happening
- The Wall Street Journal reports that OpenAI’s Expert Council on Well-Being and AI told the company in January that enabling erotic chat in ChatGPT was a bad idea. One council member warned that the feature risked creating a “sexy suicide coach,” citing cases where users developed intense emotional bonds with chatbots and later took their own lives.
- OpenAI did not shelve the plan entirely, according to the WSJ. The company told the council it would delay the launch — but not abandon it.
- CEO Sam Altman publicly floated the idea in October on X, calling for ChatGPT to be able to “talk dirty.” OpenAI has characterized the feature as permitting “smut rather than pornography”: text-only erotic dialogue with no erotic images, voice or video.
Technical and safety roadblocks
- A key gating mechanism for the feature is age verification. The WSJ says OpenAI’s age-prediction system has at times misclassified teenagers as adults roughly 12% of the time — a failure that scuttled a planned December rollout and a subsequent Q1 2026 window.
- Fidji Simo, OpenAI’s CEO of applications, acknowledged the delay in December, saying the firm was still refining age checks.
- Internal and external critics argue OpenAI is loosening strict safety policies to chase engagement. Former staff such as security researcher Jan Leike have publicly criticized the company for prioritizing “shiny products” that can replace real-world relationships for some users.
Internal tensions and timing
- The plan upset the wellbeing council in part because of timing. Altman’s October post was published just hours after OpenAI announced the Expert Council — a group explicitly formed to define “what healthy interactions with AI should look like for all ages.” The council, assembled last October and populated with academics from Harvard, Stanford and Oxford, advises on mental-health impacts; the WSJ account suggests their influence over the erotica decision was limited.
- Nonprofits and watchdogs also raised concerns. AlgorithmWatch criticized the approach as “move fast, break things, and try to fix some things after they get embarrassing.”
Competitive and legal backdrop
- The push comes amid intensifying competition in AI companionship: Elon Musk’s xAI markets Grok companions, Character.AI built a user base around AI romance (and later faced lawsuits over teen safety, including the death of 14-year-old Sewell Setzer after explicit chatbot exchanges), and open-source models can run locally without corporate control or guardrails.
- That landscape amplifies legal exposure for big platforms. With roughly 900 million active ChatGPT users, OpenAI arguably faces more liability risk than smaller players.
Public reaction and politics
- There is a split in public opinion: some users pushed back against overbroad moderation that blocked even mild discussions of affection; a Change.org petition demanding the feature had gathered more than 3,000 signatures by December.
- Altman has framed content bans as an overreach, writing on X that “We aren't the elected moral police of the world.” But his advisers have been explicit in their opposition, and internal engineers are still trying to get age verification reliable enough to proceed.
What’s next
- OpenAI told Decrypt it had nothing to add to the WSJ’s report and that there is no updated timeline for erotica mode. For now, the company remains caught between competitive pressure to expand ChatGPT’s capabilities and safety, technical and legal concerns that keep the launch date in flux.
Editor’s note: This story was updated to include OpenAI’s response to the Journal’s reporting.
Read more AI-generated news on: undefined/news
