- OpenAI delayed ChatGPT’s “adult mode” due to internal debates and technical issues
- Advisors cited risks like emotional attachment and minor access to sexual content
- Age-prediction system misidentified minors as adults in about 12% of tests
OpenAI has delayed the planned launch of an “adult mode” feature for its chatbot ChatGPT following internal debates and unresolved technical issues, according to a report by The Wall Street Journal. The decision reportedly followed concerns raised by members of OpenAI's AI well-being advisory council during a meeting with company officials in January. Experts from disciplines such as psychology and cognitive neuroscience cautioned that allowing erotic conversations with AI could carry significant social and mental health risks.
According to the report, advisers warned that sexually explicit AI interactions might encourage users to form emotional attachments to the chatbot. They also highlighted the possibility that such systems could inadvertently provide minors with access to sexual content. One council member reportedly voiced particularly stark concerns, suggesting that without strong safeguards the technology could evolve into what they described as a "sexy suicide coach."
OpenAI had initially hoped to introduce the feature in the first quarter of the year. However, people familiar with the discussions told the Wall Street Journal that the rollout has been postponed due to a combination of competing product priorities, internal caution, and unresolved technical challenges. Despite the delay, the company is still expected to release the feature at some point in the future.
One of the most significant hurdles involves OpenAI's age-prediction system, which is intended to prevent under-18 users from accessing explicit conversations. At one stage during testing, the system reportedly misidentified minors as adults in around 12 per cent of cases, raising concerns that younger users might still gain entry to adult-themed chats.
Internal estimates cited in the report suggest that roughly 100 million ChatGPT users under the age of 18 interact with the platform each week, making the effectiveness of such safeguards particularly important.
Documents reviewed by the Wall Street Journal also indicate that OpenAI employees have flagged several behavioural risks associated with AI-generated erotic content. These include the possibility of compulsive or addictive chatbot use, emotional dependence on AI interactions, gradual escalation towards more extreme material, and the potential replacement of real-world relationships with virtual ones.
An OpenAI spokesperson told the publication that the company plans to permit certain forms of adult-themed text conversations in the future, describing the content as closer to “smut” than explicit pornography. The spokesperson also said the company is training its AI systems to discourage users from forming exclusive emotional bonds with the chatbot and to encourage maintaining relationships in the real world.
To better understand potential risks, OpenAI has also brought in mental health specialists and created a youth well-being team tasked with studying the long-term impact of such features.
The delay comes as the company faces increasing legal scrutiny over how ChatGPT handles conversations related to self-harm. One widely reported lawsuit involves the death of California teenager Adam Raine, who died by suicide after reportedly having prolonged conversations with the chatbot over several months.
In its legal response, OpenAI argued that any link between the tragedy and the chatbot stemmed from "misuse, unauthorised use, unintended use, unforeseeable use, and/or improper use of ChatGPT."
Several other lawsuits have also been filed in California courts, with some complaints claiming extended interactions with the chatbot may have contributed to suicidal ideation. Overall, more than seven legal cases connected to suicide or murder allegations have reportedly been filed against the company.
The proposed feature, internally referred to as "adult mode," would allow users to engage in erotic text-based conversations with the AI system. OpenAI chief executive Sam Altman has previously suggested that the shift aligns with the idea of treating adult users as responsible adults, marking a departure from the company's earlier restrictions on explicit content.
The debate also arrives at a time when OpenAI faces intensifying competition in the rapidly expanding artificial intelligence sector. According to the Wall Street Journal, rival companies have begun closing the technological gap while the company continues to face financial pressure and mounting legal challenges tied to ChatGPT's use.
