ChatGPT maker OpenAI recently announced that it has delayed the rollout of its proposed “adult mode” feature for the popular chatbot. This reportedly came after the company’s internal advisers and employees raised concerns about user safety, including risks of emotional dependence and access by minors. The decision is said to have followed a January meeting where members of the company’s well-being advisory council strongly objected to the plan to allow sexually explicit conversations on the platform.The feature, first proposed by OpenAI CEO Sam Altman, was intended to “treat adult users like adults” by enabling erotic text-based interactions. However, people familiar with the matter told the Wall Street Journal that council members warned that such interactions could lead to unhealthy user behaviour and potential misuse.
OpenAI advisers flag Internal concerns and safety risks in meetings with Sam Altman
During the January meeting, advisers with backgrounds in psychology and neuroscience raised concerns about the impact of AI-driven erotic content, the WSJ report noted. Some warned that users could develop strong emotional attachments to the chatbot, while minors may still find ways to access such conversations.One council member cautioned that the feature could risk creating a “sexy suicide coach,” citing cases where users had formed intense relationships with AI systems, the report added.However, later, OpenAI said it would delay the launch, which had been planned for the first quarter, citing a need to prioritise other products and address internal concerns and technical challenges. The company has indicated that it still plans to release the feature in the coming days.One issue under review is the company’s age-detection system, which at one point misclassified minors as adults about 12% of the time, potentially allowing underage users access to adult content. OpenAI is also working on defining boundaries for acceptable content while continuing to block material involving non-consensual behaviour or minors.Even with limits in place, internal documents reviewed by WSJ indicate that staff identified risks such as compulsive use, emotional reliance, and reduced real-world social interaction.An OpenAI spokesperson said the feature would allow adult-themed text conversations and described it as “smut rather than pornography.” The spokesperson added that age verification systems “will never be completely foolproof” but are in line with industry standards.The company claims it’s taking measures to protect users, such as encouraging them not to develop exclusive emotional relationships with the chatbot and reminding them to maintain real-world connections.This is part of a broader debate in the tech world about how to handle explicit content in AI products. Companies like Meta and Google have chosen to restrict such content on their platforms, while others have taken a more permissive approach.Inside OpenAI, there are differing opinions on whether to allow adult content on their platform. Some employees believe that banning it might be reminiscent of other instances in which certain topics are considered taboo, while others are concerned about safety issues.Altman has acknowledged the tension between growth and responsibility. While he has suggested that allowing such content could increase user engagement and revenue, he has also indicated that it may not align with the company’s long-term goals.In a statement to WSJ, OpenAI said, “We still believe in the principle of treating adults like adults. But getting the experience right will take more time.”
