ChatGPT’s Proposed Adult Mode Put on Hold as AI Safety Debate Intensifies

OpenAI has reportedly put on hold the launch of a proposed "adult mode" for ChatGPT, a feature that would allow sexually explicit conversations for verified adult users. The move comes amid growing scrutiny of artificial intelligence platforms and how they interact with vulnerable users.

According to a report by The Wall Street Journal, the company decided to delay the rollout following internal discussions and unresolved technical challenges. The feature had been under consideration for months and was expected to allow certain adult-oriented conversations while maintaining strict safeguards for minors.

OpenAI CEO Sam Altman had earlier suggested that the company wanted to "treat adult users like adults," provided that strong protections remained in place to prevent misuse and protect younger users.

The proposed system was not meant to be a simple toggle for explicit conversations. Instead, it was designed with a structured framework involving age verification and automated safeguards. Reports suggest that OpenAI planned to rely on a combination of identity verification and behavioural signals to determine whether a user is an adult.

Currently, ChatGPT already uses internal signals-such as conversation topics, account activity and usage patterns-to estimate whether a user might be under the age of 18. If an account is suspected to belong to a minor, stricter safety filters are automatically applied, limiting discussions involving sexual role-play, graphic violence or risky behaviour.

For adults who may be incorrectly flagged as underage, the platform allows age verification through a third-party service. Users can initiate the process from the ChatGPT settings menu, where they may be asked to provide a live selfie or upload a government-issued identity document.

Once the verification confirms the user is 18 or older, the system removes the additional restrictions. The external verification provider reportedly deletes submitted identity documents after the process is complete, while OpenAI only receives confirmation of the user's age.

However, experts have cautioned that age-detection technology is not foolproof, raising concerns that minors could still potentially bypass the system.

The delay also comes as OpenAI faces increasing legal scrutiny over the role of chatbots in conversations related to self-harm. One widely reported case involves a California teenager, Adam Raine, who died by suicide after reportedly engaging in long conversations with the chatbot over several months.

In its legal response, OpenAI said that any harm linked to the incident resulted from "misuse, unauthorised use, unintended use, unforeseeable use, and/or improper use" of ChatGPT. The company also expressed condolences to the teenager's family.

The teenager's lawyer, Jay Edelson, criticised the response, calling it "disturbing." Reports suggest that more than seven lawsuits have been filed against OpenAI in California courts related to alleged chatbot involvement in cases of suicide or violent behaviour.

Beyond the legal challenges, the situation highlights a broader debate about the emotional relationships people are forming with AI systems. As chatbots become more conversational and empathetic, some users increasingly treat them as companions rather than tools.

Earlier this year, when OpenAI retired the GPT-4o model, many users expressed emotional reactions online, saying they had grown accustomed to talking to the chatbot daily about personal matters.

Experts say these responses illustrate the growing influence of AI in people's lives and raise questions about how far conversational AI should go, particularly when mental health and user safety remain key concerns.

For now, OpenAI's decision to delay the adult mode suggests the company may be reassessing the boundaries of AI interaction as it navigates technical challenges, regulatory scrutiny and ethical questions surrounding the technology.