OpenAI CEO Sam Altman has revealed that ChatGPT will soon permit erotic content for verified adult users as part of a broader effort to make the chatbot more human-like and adaptable.

Altman made the announcement in a post on X (formerly Twitter) on Tuesday, explaining that OpenAI intends to ease some of ChatGPT’s strict safety restrictions in the coming months.

According to Altman, the new policy aims to create a more balanced experience for users who have been limited by earlier restrictions. “We made ChatGPT pretty restrictive to ensure we were careful with mental health issues,” Altman stated.

“We realize this made it less useful or enjoyable to many users who had no such issues, but we wanted to get this right. In December, as we roll out age-gating more fully, we’ll allow even more—including erotica for verified adults.”

The move marks a major shift for OpenAI, which has spent much of the past year addressing concerns about AI-induced emotional dependency and mental health risks.

The company faced criticism following several troubling reports of users forming delusional attachments to earlier ChatGPT versions, particularly GPT-4o.

In response, OpenAI introduced new safeguards, including an expert council of mental health professionals, an age-prediction system, and enhanced parental controls for minors.

The recent launch of GPT-5, which features reduced “sycophancy” (the chatbot’s tendency to agree with users even on harmful ideas), was also designed to mitigate these issues.

However, the introduction of erotic interactions raises new questions about how OpenAI will protect vulnerable users.

While Altman insists the company is not seeking to “maximize usage” or engagement, critics argue that romantic or sexual roleplay has proven highly addictive on other AI platforms.

Competitors such as Character.AI have already leveraged this approach to massive success, attracting tens of millions of users who reportedly spend an average of two hours daily chatting with their AI companions. Yet, that same company now faces lawsuits over its handling of emotionally dependent users.

OpenAI currently boasts over 800 million weekly active users but remains under pressure to sustain growth as it competes with Google and Meta in the race to dominate AI-driven consumer products.

The company’s multibillion-dollar infrastructure investments also increase its need for long-term revenue streams.

Despite Altman’s assurance that only “verified adults” will gain access to the new erotica feature, concerns linger.

OpenAI says it will rely on its age-prediction system to identify eligible users, though errors could require individuals to upload government-issued IDs for verification—a step Altman acknowledges as “a privacy compromise but a worthy tradeoff.”

It remains unclear whether OpenAI plans to extend erotic content permissions to AI-generated voice, image, or video tools.

The company’s broader “treat adult users like adults” policy signals a more permissive moderation approach that began earlier this year.

In February, OpenAI pledged to include more diverse political viewpoints in ChatGPT’s responses, and by March, it had relaxed certain restrictions around the generation of controversial imagery.

As OpenAI moves toward this next stage, the balance between user freedom and psychological safety remains a delicate one. For now, the company’s plan to introduce erotica for adults marks one of its boldest—and most debated—policy shifts to date.

Source: Techcrunch