OpenAI will be making changes to ChatGPT over the next few months regarding the moderation policy. As Sam Altman, CEO of the company, announced in a message through X, the chatbot will implement an age control to allow a more adult interaction with the tool, even allowing erotic content.

The movement responds, according to the manager, to the OpenAI principle of “treat adult users like adults“He assures that it has been done to ChatGPT”quite restrictive“with the intention”to be careful with mental health problems”, an approach that, as Altman explains, has made the assistant “less useful and enjoyable for many users”.

The head of the company assures that it now has new tools and will be able to relax the restrictions safely in most cases. “In a few weeks, we plan to release a new version of ChatGPT that allows users to have a personality more like the one they liked from 4o“, Altman has noted. “If you want your ChatGPT to respond in a very human way, use a lot of emojis or behave like a friend, it should do so”.

That is why, starting in December, ChatGPT will allow verified adults to access erotic content in their conversations with the chatbot. The option will be the result of the deployment of the new age control system that will be able to automatically identify when interacting with an underage user. The tool will then apply certain restrictions and block unsuitable content to offer the most appropriate experience for each user’s age.

In this sense, interaction with ChatGPT starting in the coming months will be more malleable, since it will facilitate the customization of the assistant’s tone, a function that many users have missed between the GPT-4o and GPT-5 versions.
“It’s about time… ChatGPT used to seem like a person you could actually talk to, but then it turned into a compliance bot. If you can make it fun again without losing the barriers, it is a great achievement. People don’t want chaos, just authenticity“, a user commented in response to Sam Altman’s publication.

Many users have taken the opportunity to express their opinion regarding the changes to the models made by OpenAI. From their comments it is clear that, in the transition from GPT-4o to GPT-5, many of them lost value and that the company is not really addressing their comments and needs.

“You are misidentifying the core value that users saw in GPT-4o. It wasn’t the tone. It wasn’t the emojis. It wasn’t kindness. It was the model’s ability to interpret disordered, informal language with multiple emotional nuances with coherent responses, adaptive to the context and aligned with the objectives, without the need for the user to be precise.“, explains one user to Altman. “If you try to recreate 4o by modifying superficial behaviors, you are solving the wrong problem”.

Altman’s message aims to convey that, three years after the launch of the tool, OpenAI has been able to develop more robust mechanisms to guarantee security. However, many question how effective these methods will really be and how they will adapt to the legal differences in each jurisdiction. Consequently, many users see in these types of movements the need to strengthen regulation for this type of technology.

OpenAI’s decision is also understood on a competitive level. And it could be losing users in favor of other services, such as Grok from xAI, which in recent months have already experimented with artificial intelligences with more marked personalities, called Companions, and offer flirtatious or irreverent tones.