In February 2025, OpenAI announced its intention to “give more freedom” to its users by allowing ChatGPT to have “spicy” discussions with the chatbot. An update that was a long time coming, since according to The Verge, Sam Altman’s company will soon open the floodgates in terms of “mature applications.”
The goal is to allow discussions about “soft porn” once the age verification system is launched. However, it will not be possible to have such discussions involving a minor (or to be accessed by a minor user).
“Spicy” discussions that open the door to abuse
This should result in the emergence of chatbots designed specifically for this type of discussion, such as Grok, which offers virtual avatars that can make sexually explicit comments based on user requests. His “character” is based on dark romance characters, as Elon Musk wanted.
The possible safeguards around this new rule remain to be known, in particular with regard to deepfakes, which include, for example, film personalities. At a time when OpenAI is under particular scrutiny, the arrival of “mature” discussions risks posing a problem for regulators.
In fact, ChatGPT has been accused for several months of not doing enough to prevent users from using it inappropriately, with the problems this can create. More and more people take the chatbot as a confidant or ask it for advice on health problems. But OpenAI’s generative AI is far from trustworthy, particularly on mental health issues, and sometimes even gives bad advice to its users.
In this context, many users and associations criticize OpenAI for its opacity.
Additionally, as Futurism points out, modifying a large language model requires a lot of work and patience. Since they have been trained for years, their behavior after a potential patch may not change, or at least not immediately.
Source: BFM TV
