Adult Mode in ChatGPT Reportedly Put on Hold
According to a Financial Times report, the San Francisco-based AI giant has decided to delay Adult Mode in ChatGPT indefinitely to focus on its core products, specifically Codex and the expansion of the chatbot. One of the factors behind the latest delay is also concerns about the effect of a sexualised AI on the larger society.
Citing unnamed people familiar with the matter, the publication claimed that OpenAI is also holding internal discussions on whether the project should be scrapped entirely. Adult Mode was first announced in October 2025; however, the company delayed it to research the effects of sexual chats with an AI chatbot on individuals. At the time, it admitted that there was no empirical data on the same.
A second delay came earlier this month when a report claimed that OpenAI was shifting its focus to other higher-priority projects. Now, the FT report claims that the decision was also influenced by concerns raised by employees and investors. While the company was targeting increased engagement and usage growth, the staff reportedly highlighted the ethical and reputational risks of such a feature.
The main concern seems to be emotional dependence. With an AI chatbot flirting and engaging in romantic conversations with a user, there is a higher chance that the user will develop an emotional connection, which can lead to unhealthy attachments. A couple of investors have reportedly highlighted these risks and compared them to the relatively small upside for the business.
Another challenge OpenAI is facing is a technical one. As per the report, the AI giant is struggling to train its AI models to generate explicit content after spending a long time training them to avoid such conversations. Training data is reportedly also a concern, as such datasets can contain unethical or illegal content, removing which could be more costly and resource-intensive.
Additionally, the feature was slated to be released alongside the company’s age verification system when Adult Mode was first announced. However, the report claims that OpenAI’s age verification system currently has an error rate of 10 percent. This could lead to ChatGPT providing access to the explicit feature to minors, a mistake OpenAI will be wary of making after facing multiple lawsuits.
