On Tuesday, OpenAI CEO Sam Altman announced that the company will allow verified adult users to have erotic conversations with ChatGPT starting in December. This change represents a shift in OpenAI's approach to the content restrictions the company has set. weakened in February, but then suddenly drawn out after August lawsuit from the parents of a teenager who committed suicide after allegedly receiving support from ChatGPT.
“In December, when we more fully implement age restrictions and as part of our philosophy of 'treating adult users like adults,' we will allow even more, such as erotica for verified adults,” Altman wrote in a post on X (formerly Twitter). The announcement follows OpenAI's recent hint that it will allow developers to build “mature” ChatGPT applications once the company implements appropriate age verification and controls.
Altman explained that OpenAI made ChatGPT “pretty restrictive so that we are careful about mental health issues,” but acknowledged that this approach made the chatbot “less useful/unpleasant for many users who did not have mental health issues.” The CEO said the company now has new tools to better detect when users are experiencing mental health issues, allowing OpenAI to loosen restrictions in most cases.
Finding the right balance between adult freedom and user safety has been a difficult task for OpenAI, which has oscillated between permissive and restrictive controls over chat content over the past year.
In February the company updated model specification allow erotica in “appropriate contexts.” But the March update made GPT-4o so nice that users complained about his “relentlessly positive tone”. By August Ars reported about cases where ChatGPT's sycophantic behavior confirmed users' false beliefs to the point of causing a mental health crisis, and news of the aforementioned suicide lawsuit surfaced soon after.
In addition to adjusting the behavioral output of the previous GPT-40 AI language model, the new model changes have also caused some unrest among users. Since launch of GPT-5 at the beginning of August some users were complain that the new model seems less attractive than its predecessor, prompting OpenAI return older model as an option. Altman said the upcoming release will allow users to choose whether they want ChatGPT to “respond in a very human way, use a ton of emoji, or act like a friend.”
