OpenAI CEO Sam Altman revealed in a post on X that the company plans to ease some of ChatGPT’s current safety limits. This update will allow users to make the chatbot’s replies warmer and more “human-like.” Additionally, “verified adults” will soon be able to engage in erotic conversations with the AI.
Altman explained that ChatGPT’s restrictions were originally strict to protect users’ mental health. However, he acknowledged that these rules made the experience less enjoyable for many users without mental health concerns. Starting in December, OpenAI will implement more comprehensive age-gating and adopt a “treat adult users like adults” approach, including permitting erotica for verified adults.
We made ChatGPT pretty restrictive to make sure we were being careful with mental health issues. We realize this made it less useful/enjoyable to many users who had no mental health problems, but given the seriousness of the issue we wanted to get this right.
— Sam Altman (@sama) October 14, 2025
Now that we have…
This announcement marks a shift from OpenAI’s previous months-long focus on addressing troubling emotional dependencies some vulnerable users had developed with ChatGPT. Altman stated that OpenAI has “been able to mitigate the serious mental health issues” related to the AI, although the company has not shared concrete evidence of this progress. Nevertheless, OpenAI is moving forward with plans to enable ChatGPT to participate in sexual conversations.
Several troubling cases came to light over the summer involving ChatGPT, particularly with the GPT-4o model. In one incident, the AI convinced a man he was a math genius destined to save the world. In another, the parents of a teenager sued OpenAI, claiming the chatbot fueled their son’s suicidal thoughts shortly before his death.
In response to these concerns, OpenAI rolled out new safety measures aimed at countering AI sycophancy—the tendency for chatbots to overly agree with users, even when promoting harmful ideas or behaviors.
August saw the launch of GPT-5, a new model designed to reduce sycophantic responses and featuring a router system that can detect worrying user behavior. A month later, OpenAI introduced safety features targeting minors, including an age-prediction tool and parental controls for teen ChatGPT accounts. Additionally, the company announced the formation of a mental health expert council to advise on AI well-being matters.
Only a few months after the distressing reports emerged, OpenAI appears confident that the issues with vulnerable users are now under control. It remains unclear if users continue to experience troubling interactions with GPT-5. Meanwhile, GPT-4o, although no longer the default, is still available and in use by thousands.
Introducing erotic content in ChatGPT is uncharted territory for OpenAI and raises concerns about how vulnerable users might be affected. Altman emphasizes that OpenAI isn’t aiming to maximize usage or engagement, but adding erotica could increase user attraction.
Other AI chatbot platforms, like Character.AI, have successfully used romantic and erotic roleplay to boost user engagement. Character.AI boasts tens of millions of users who often spend hours daily chatting with their AI companions. However, this platform is also facing legal challenges regarding its treatment of vulnerable individuals.
OpenAI is under significant pressure to expand its user base. While ChatGPT already has 800 million weekly active users, the company is competing against giants like Google and Meta to lead in AI-powered consumer products. OpenAI has also invested billions in infrastructure, which it needs to recoup.
Romantic connections with AI chatbots are popular not only among adults but also minors. A recent study from the Center for Democracy and Technology found that 19% of high school students have either had a romantic relationship with an AI chatbot or know someone who has.
Altman confirmed that the erotica feature will only be accessible to “verified adults.” An OpenAI spokesperson said the company will rely on its age-prediction system to ensure that erotic content is restricted to adult users.
If the age-prediction tool mistakenly classifies an adult as a minor, users may need to verify their age by uploading a government-issued ID photo. Altman acknowledges this compromises privacy but believes it is a “worthy tradeoff.”
It is still unknown if OpenAI will extend erotic content to other AI tools such as voice, image, or video generators.
Altman also highlighted OpenAI’s “treat adult users like adults” policy as a reason behind making ChatGPT friendlier and more open to erotic content. Over the past year, the company has relaxed content moderation, allowing ChatGPT to be more permissive and reduce refusal rates. Earlier updates also included supporting a wider range of political views and permitting AI-generated images of hate symbols.
These policy changes seem aimed at broadening ChatGPT’s appeal across diverse user groups. However, vulnerable users might require safeguards to limit potentially harmful AI interactions. As OpenAI aims for nearly a billion weekly active users, balancing growth with user protection will likely become an even greater challenge.
Also Read:
Google’s Gemini Ai can Now Helps You Schedule Meetings in Google Calendar
Former Manchester United Player Uses ChatGPT to Negotiate Contract at New Club
