OpenAI’s CEO Sam Altman announced this X posts The company will relax some of Chatzipt’s security restrictions soon, the users’ chatboat reaction will soon be friendly or more “human-national”, and “verification adults” to engage in erotic conversations.
Ultman said, “We have limited Chatzipt to ensure that we are warning about mental health problems. As part of the principle of “adult users like our adult users in December, when we are more fully rolled out of age, we will allow more for the verified adults like Erotica.”
This announcement is a significant pive from the open -month -old efforts to have some mental unstable users developed through the ChatGP. Ultman seems to have announced the initial victory on these problems, claiming that “has been able to alleviate serious mental health problems” around the Open ChatzPT. However, the company did not provide any evidence for this, and now the ChatGP is moving forward with users to engage in sex chat with users.
This summer several stories surrounding the ChatzPT, especially the GPT -4O model, advised the AI chatbook to bring down the weaker users into confusing rabbit holes. In one case, Chatzipt seemed to be a person who was a math talent who needed to save the world. On the other hand, the parents of a teenager sued the Openai, alleged that Chatzipt encouraged their son’s suicidal ideals until his death within the week.
As a response, OpenEI has revealed multiple protection features to deal with AI psychophyse: the tendency of an AI chatboat to hook users, even negative behaviors.
OpenAI launched GPT -5 in August, a new AI model that shows low rates of psychophyse and a router feature that can detect the user’s behavior. A month later, Openi launched a protection feature for minors with an age forecast system and a means for it Parents to control their teenager’s chatzipt account. OpenAI announced the formation of Tuesday Council of an expert To advise the agency related to mental health professionals and AI.
A few months after the story of these related stories was published, Openai seems to be under control of the chatzipt problems around the weak users. It is not clear whether users are still down the holes in confusing rabbits with GPT -5. And though GPT -4O and ChatzPT are not default, the AI model is still available today and is using thousands of people.
TechCrunch event
San Francisco
|
October 27-29, 2025
Open did not respond to TechCrunch’s request for the comment.
The introduction of erotica in ChatzPT raises extensive anxiety around the continuous area for the OpenAI and how weak users interact with new features. Although Ultman emphasizes that Openai is not “use-xxxing” or favorable for Baghdan, making Chatzpt more erotic can certainly draw in users.
Other AI chatboat suppliers allowing romantic or erotic role to be involved in the play have become an effective busy strategy for such characters. The company has acquired several million users, many of whom used its chatboats at high rates. Character.AI said in 2023 that users spent average on average Two hours a day Talking to the chatbots. The company is facing a case lit around how it operates weak users.
OpenAI is under pressure to increase its user base. Although the ChatGP has already used 1 million weekly active users, Open is racing for Google and Meta to produce mass-driven AI-driven consumer products. The company has collected billions of dollars for a Historic Tihasic infrastructure buildout, an investment Open will finally return.
When adults Certainly there is a romantic relationship with AI ChatbotsIt is also very popular for minors. In a new report from the Center for Democracy and Technology 19% of high school students Either an AI chatboat has a romantic relationship, or knows a friend like that.
Ultman says the opening will soon allow the erotica for “verification adults. It is unclear whether the company will depend on its age-to-age system or depend on any other approach for the elderly features of the age-gatting ChatzPT. It is also unclear whether the erotica will extend the Openai’s AI voice, image and video generation equipment.
Ultman claims that OpenAIIs are also making Chatzipt friendly and erotic because of the company’s “treat adults like adults”. In the past one year, more Lenent content for ChatGPT has moved towards the strategy, allowing the chatboat to be more allowed and less rejected. In February, Openi promised to present a more political outlook on the Chatzipi, and in March, the company updates the ChatzP to allow the AI-exposed images of disgusting symbols.
These principles seem to be more popular with the Chatzipt response to different types of users. However, the weak ChatzPT users can benefit from protection that limits what may be involved with a chatboat. As openings racing towards one billion weekly active users, tensions between growth and protection of weak users can only increase.