Why OpenAI is Putting ChatGPT Adult Mode on Ice

Controversy and widespread concern means ChatGPT’s sexually explicit ‘adult mode’ has been shelved. 

The feature – reportedly known internally as Citron mode – had been positioned as a controlled way for consenting adults to engage in explicit conversations using AI. 

But what began as an experiment in digital intimacy has ended in a strategic withdrawal, shaped by mounting unease inside and outside the company.

The announcement to indefinitely drop adult mode comes as Meta and YouTube were found liable in a landmark social media addiction trial.

First outlined in late 2025 by CEO Sam Altman, the concept rested on a simple premise: that adult users should be permitted greater conversational freedom, provided robust safeguards were in place. 

A previous X post from Sam read: “In December, as we roll out age-gating more fully and as part of our ‘treat adult users like adults’ principle, we will allow even more, like erotica for verified adults.”

However, translating the ‘age-gating’ principle into practice proved far more complex, with a New York Post report noting that the misclassification rate – where a minor was classified as an adult – was 12%.

The Code Red roadmap

Initially slated for release at the beginning of 2026, the erotic feature was delayed after Sam reportedly issued a company “Code Red” asking employees to focus all attention on AI models to combat rivals. 

The strategic priority shift has also seen the business scrap its popular video generation platform Sora and rescind a billion-dollar partnership with Disney. 

As OpenAI gears up for its IPO, CEO of Applications Fidji Simo says the organisation cannot afford to be distracted by “side quests”.

Plans for a desktop super app encompassing numerous offerings are also in the pipeline. 

The mental health angle

Concerns quickly emerged around the psychological and societal impact of erotic AI.

Advisers and researchers warned that highly-personalised chatbot interactions could foster unhealthy emotional attachments, particularly among vulnerable users. 

These concerns were not merely theoretical. The broader AI sector has already faced legal and ethical scrutiny over cases involving emotional dependency and harmful advice from chatbots.

xAI’s Grok has come under legal fire for its role in producing fake, sexualised images of women and children. 

With OpenAI's age verification feature not being completely reliable, even a small failure rate has raised the prospect of millions of minors accessing explicit interactions – a risk few technology companies can afford to take in an increasingly regulated environment.

Against this backdrop, OpenAI’s proposed adult mode began to look less like a bold innovation and more like a reputational liability.

The decision to shelve the feature was also shaped by internal resistance. Employees and investors reportedly questioned whether the project aligned with OpenAI’s long-term mission, particularly as the company prepares for deeper commercial expansion. 

At the same time, regulators and policymakers have been sharpening their focus on AI safety, especially in areas involving sexuality and vulnerable users.

This convergence of pressure appears to have forced a recalibration. Rather than pushing ahead into uncertain territory, OpenAI has chosen to pause indefinitely, citing a lack of clear evidence on the long-term effects of explicit AI interactions.

OpenAI’s retreat reflects a wider moment of reckoning for the AI industry and, for now the concept of a mainstream adult chatbot remains on hold. 

Whether it returns in a safer, more regulated form will depend on society’s evolving tolerance of intimacy mediated by machines.