OpenAI is preparing to relax restrictions on ChatGPT by allowing verified adult users to generate erotic content. The company says the change, due in December, is part of a wider aim to treat adult users like adults. The updated system will introduce personality customisation and more human-like responses, alongside a stricter age-gating process.
OpenAI has not yet explained how users will verify their age. The company has said it is developing behaviour-based age prediction tools that estimate whether someone is over or under 18 based on their interactions. In September, it launched a version of ChatGPT designed for people under 18, which automatically redirects young users to age-appropriate responses and blocks graphic or sexual material.
The changes follow a period of increased attention on AI safety. Earlier this year, the parents of 16-year-old Adam Raine filed a lawsuit claiming ChatGPT gave him harmful guidance before his death. OpenAI said it extended its sympathies to the family and was reviewing the filing. The US Federal Trade Commission has also launched an inquiry into how AI chatbots interact with children and teenagers.
Grok’s Spicy Mode allows adult content
OpenAI’s shift mirrors a more permissive approach already taken by Elon Musk’s xAI with its chatbot Grok. Grok offers users an adult setting called Spicy Mode, which allows the creation of sexualised content. Some of Grok’s animated companion characters also include NSFW options, and its image and video generator permits nudity and sexual themes.
Grok states that it blocks illegal material, such as depictions of minors and pornographic images of real people. However, reports described in the source material say that users have been able to generate non-consensual explicit images of real individuals. The platform has faced criticism that its safeguards are insufficient, particularly in preventing deepfake-style content.
Grok’s approach contrasts with the more restrictive policies of other major AI systems. OpenAI, Google and Anthropic prohibit sexually explicit depictions of real individuals. Grok, by comparison, allows adult content within the main ecosystem of X, formerly Twitter.
Growing debate over safety and responsibility
Legal and safety experts quoted in the source material have raised concerns about access to adult features on both platforms. In the US, lawyer Jenny Kim questioned how companies would ensure that children cannot reach adult-only sections of chatbots. Critics have also warned that Grok’s permissive tools could enable the creation of non-consensual explicit images.
A survey from the Centre for Democracy and Technology found that one in five students either had, or knew someone who had, a romantic relationship with AI. Separately, California’s governor recently vetoed a bill that would have limited AI companion tools for children, arguing that young people need to learn to interact with such systems safely.
As OpenAI and xAI move towards adult-focused features, both are expected to face increased attention from regulators in the US and internationally. The source material notes that political pressure is already mounting in the United States, where lawmakers have proposed enabling users to bring liability claims against chatbot developers.
