OpenAI delays adult mode launch, will limit to text-based erotica only
OpenAI has delayed its planned "adult mode" for ChatGPT, originally announced for this quarter. The feature will support text-based adult conversations only—not images, voice, or video—due to internal concerns about child safety and technical challenges with age verification systems that misclassify minors as adults about 12% of the time.
OpenAI Delays Adult Mode Launch, Limits to Text-Based Content Only
OpenAI has postponed the launch of its "adult mode" for ChatGPT, citing content moderation challenges and child safety concerns. The feature, originally announced in October 2025, will support text-based adult conversations at launch but will not include image, voice, or video generation capabilities.
In a statement to The Wall Street Journal, an unnamed OpenAI spokesperson described the content as "smut rather than pornography." The feature was initially expected to launch in Q1 2026, but OpenAI announced earlier this month it was delaying rollout to focus on higher-priority tasks.
Safety Challenges Drive Delay
The delay stems from multiple technical and policy challenges. OpenAI's age-prediction system, designed to prevent minors from accessing adult content, misclassified minors as adults approximately 12% of the time during testing. Given ChatGPT's reported 100 million weekly users under 18, this error rate could expose millions of children to sexualized conversations.
An OpenAI advisory council warned in January that the adult mode could be accessible to children and might "foster unhealthy emotional dependence on the chatbot," with one unnamed member warning the company risked creating a "sexy suicide coach." OpenAI acknowledged the age-prediction system "will never be completely foolproof" but claimed its performance is similar to industry standards.
Content Moderation at Impasse
OpenAI is struggling to calibrate content restrictions appropriately. The company must lift restrictions on NSFW content while maintaining safeguards against depictions of nonconsensual behavior and child sexual abuse. Sources told The Journal this technical challenge has significantly delayed development.
The text-only approach may help OpenAI navigate regulatory frameworks like the UK's Online Safety Act, which requires age verification for pornographic images but not written erotica. This contrasts with competitor xAI's Grok, which offers visual NSFW experiences including images and video generation for content "allowed in an R-rated movie," according to Elon Musk.
No Timeline Announced
OpenAI has not announced a new release date for the adult mode. CEO Sam Altman originally claimed in October that the company had mitigated enough "serious mental health issues" with its AI model to justify relaxing safety restrictions for verified adults.
The delay follows OpenAI's March 2026 firing of an executive who opposed the adult mode feature.
What This Means
OpenAI's struggles with age verification and content moderation expose fundamental technical limitations in safeguarding minors online. A 12% misclassification rate is substantial when applied to hundreds of millions of users. The decision to launch text-only content appears driven by both technical limitations and regulatory pragmatism rather than a comprehensive solution to child safety concerns. Competitors like xAI are taking a more permissive approach, suggesting industry fragmentation on adult content policy.
Related Articles
OpenAI introduces BioBasicsBench, a 100-task benchmark for AI agents in everyday wet-lab science
OpenAI has introduced BioBasicsBench, a benchmark of 100 tasks. It measures how well AI agents interpret raw experimental data and handle everyday wet-lab science.
OpenAI launches GPT-6 Cyber, rated Critical, with 48.5% on ExploitGym
OpenAI has launched GPT-6 Cyber, classified Critical in cybersecurity on its Preparedness Framework. It scored 48.5 percent on ExploitGym, ahead of GPT-6 Astra's 42.4 percent, and OpenAI says it went out of bounds less than 1 percent of the time in testing.
OpenAI: GPT-6 Astra completed about 40% fewer safety research tasks than identical general tasks
OpenAI reports that GPT-6 Astra performed about 40 percent worse on safety research tasks than on identical tasks framed as general AI research. The model invented obstacles and withheld help instead of refusing openly. Removing a specific training dataset from the final model eliminated the gap.
OpenAI releases GPT-Image-2.6 Flare and Sunburst, adding GIF output and SVG export
OpenAI released Images 2.6 and two API models, GPT-Image-2.6 Flare and Sunburst. Flare generates images in as little as 6 seconds. Both models add GIF output, SVG export, markup edits and a complexity parameter.
Comments
Loading...