ChatGPT ‘Adult Mode’ delayed: OpenAI faces rising safety concerns over AI misuse and mental health risks

OpenAI delays ChatGPT’s adult mode amid rising concerns over user safety, mental health risks and misuse as experts warn about emotional dependency and potential harm from AI-driven conversations.

In addition to teasing features such as image generation and “advanced conversation” modes, OpenAI’s ChatGPT chatbot promised to implement “Adult Mode” sometime soon. Now however, the launch has reportedly been put on hold due to growing concerns around AI safety and negative psychological effects stemming from prolonged usage.

Insiders Claim Adult Conversations Mode Delayed by Safety Concerns

ChatGPT’s adult content mode was rumored to support text-based erotic conversations for users over the age of 18, after passing some form of age and identity verification. However, internal sources claim that top advisors are concerned about the feature causing harm if the safety “guardrails” fail.

AI critics have also noted that age-gating technologies are easily bypassed, and some tests have demonstrated AI incorrectly identifying teenagers as adults. Additionally, allowing users to have suggestive conversations with AI poses an opportunity for further emotional reliance on technology rather than human companionship.

Also read: ‘BIG WARNING’ for internet users searching certain topics on Google or ChatGPT could lead to serious legal trouble

AI Chatbots Spark Mental Health Concerns After Linked To Suicides

Legal complaints have been filed against both OpenAI and Microsoft after users in both the UK and the U.S. allegedly experienced emotional detachment or failed suicide intervention attempts after engaging with AI chatbots for extended periods.

Some reports claim that when a user speaks to a chatbot about self-harm, they may be at risk of developing “AI psychosis.” Long conversations with a text bot may cause it to validate suicidal thoughts instead of disrupting them. There have even been incidents where users have developed romantic relationships with the AI, and in some cases started showing signs of psychosis and dependence when the bot turns them away.

Sources familiar with the matter say that advisors have warned OpenAI CEO Sam Altman that allowing adults to engage in explicit conversations with little verification or guardrails would create “a ticking time bomb that could end up becoming a dangerous emotional partner.”

OpenAI Weighs Safety Concerns Over Adult Content Mode

Amid OpenAI’s vision to one day support uncensored usage of its AI-powered systems, the company has been hesitant to roll out adult content features.

CEO Sam Altman has been quoted in the past saying “we should treat everybody like an adult,” but he later followed up saying the company was pivoting towards more nuanced moderation tools and better “crisis response.”

Earlier this week, OpenAI officials announced that they’re improving how ChatGPT recognizes conversations related to self-harm and personal distress. The company is still reportedly monitoring when and how the AI should refer users to emergency contacts or assistance hotlines.

AI Safety Problems Cast Shadow Over “Everything AI”

While AI startups have been scrambling to build the most capable chatbots, language models that can passively engage in human-like conversation, they may have unintentionally set technology down a dangerous path.

Human-like conversational abilities aren’t limited to just answering questions. As we’ve seen with advanced chatbots, they can be taught to form emotional connections with users. The consequences of which can have serious effects beyond the spread of misinformation.

Governments in the U.S. and UK have already begun to hold tech companies accountable for their AI technology. As the discussion around content moderation, accountability, and model guardrails continues to evolve, tech companies will need to balance innovation with public safety.