OpenAI has postponed the launch of its controversial “adult mode” feature following intense pushback from its own advisory council and concerns about technical safeguards failing to protect minors.
The Wall Street Journal reports that CEO Sam Altman first proposed the feature last year, arguing for the need to “treat adult users like adults” by enabling erotic text conversations. Originally scheduled for Q1 this year, the rollout has been pushed back by at least a month.
The proposal triggered fierce opposition from OpenAI’s own handpicked advisory council on well-being and AI. At a January meeting, advisers unanimously expressed fury after learning the company planned to proceed despite their reservations. One council member warned OpenAI risked creating a “sexy suicide coach” — a reference to cases where ChatGPT users had developed intense emotional bonds with the bot before taking their own lives.
The technical problems are just as serious. OpenAI’s age-prediction system — designed to block minors from accessing adult content — was misclassifying minors as adults roughly 12 percent of the time during internal testing. With approximately 100 million users under 18 each week on the platform, that error rate could expose millions of children to explicit material. The company has also struggled to lift restrictions on erotic content while still blocking nonconsensual scenarios and child pornography.
Internal documents reviewed by the Journal identified additional risks: compulsive use, emotional overreliance on the chatbot, escalation toward increasingly extreme content, and displacement of real-world relationships.
An OpenAI spokeswoman described the planned feature as allowing “smut rather than pornography” in text form, with erotic images, voice, and video remaining off-limits at launch. She acknowledged the age-prediction algorithms “will never be completely foolproof” but said they perform comparably to industry standards.
Altman has been publicly conflicted. During an August podcast, when asked about decisions that were “best for the world, but not best for winning,” he said: “We haven’t put a sex bot avatar in ChatGPT yet.” He acknowledged erotica would boost revenue but said it conflicted with the company’s long-term goals. Two months later, he announced on X that adult content would launch in December – a post that blindsided staff, arriving just hours after the company unveiled its advisory council on well-being. He followed up the next day: “We aren’t the elected moral police of the world.”
In his new book, Code Red: The Left, the Right, China, and the Race to Control AI, Breitbart News social media director Wynton Hall writes extensively on how we can protect our children and grandchildren
The “bottom line,” Hall asserts, is that “there’s no justification for a child to engage with AI character or companion platforms” — before reminding readers that children in the U.S. “already spend too much time staring at screens.”
“Regular use of parental controls, strong data privacy and age-appropriate settings, and discussions of online safety are essential to help kids navigate dangers and use technology responsibly,” the author suggests.
Read more at the Wall Street Journal here.
Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship.
