Comfort talk made ChatGPT 4o a darling to some users, now even OpenAI CEO Sam Altman is worried

India Today

12 hours ago

It begins quietly enough: a voice note, a thought recorded at the end of a long day, a conversation with an always-available listener. But increasingly, that listener isn't a friend, a family member, or a therapist. It is ChatGPT, particularly ChatGPT 4o, which apparently has a personality that validates whatever its users are feeling with its sweet nothings.While this novel way of talking to an AI chatbot was increasingly getting noticed, it all came in limelight a few days ago with the launch of the ChatGPT 5. The new AI chatbot from OpenAI has a different personality and a fairly vocal group of users immediately demanded that the company bring back the 4o. Reason? They love ChatGPT 4o for what it can utter back to users and how it can validate their feelings, right or wrong, sensible or not.advertisementAcross social media platforms like Instagram and Reddit, people are now sharing how they have turned the AI chatbot into a personal sounding board.
The practice, often called 'voice journaling,' involves speaking directly to the bot, using it as both recorder and respondent. On Reddit, there is a whole thread on how many users have been asking ChatGPT for relationship advice, comfort during anxious moments, and even help processing grief. Some described it as a 'safe space' for unloading feelings they couldn't share elsewhere. Others said they journalled with ChatGPT daily, treating it almost like a life coach that never judged, interrupted, or grew tired.
It's easy to see the appeal. The bot, unlike a therapist, doesn't charge by the hour; it responds instantly, and it can appear endlessly patient. But the growing use of AI in such intimate ways has started to worry even those at the helm of its development.Sam Altman tries to nudge users awayOpenAI CEO Sam Altman has publicly warned that people should be careful before treating ChatGPT as a therapist. 'People talk about the most personal shit in their lives to ChatGPT. People use it, young people, especially, as a therapist, a life coach; having these relationship problems and (asking) what should I do?' he recently said on a podcast with comedian Theo Von.
Unlike real therapy, conversations with ChatGPT are not protected by doctor-patient or legal privilege. 'Right now, if you talk to a therapist or a lawyer or a doctor about those problems, there's legal privilege for it. And we haven't figured that out yet for when you talk to ChatGPT,' Altman acknowledged. Deleted chats, he added, may still be retrievable for legal or security reasons.advertisementPrivacy isn't the only concern. A recent Stanford University study found that AI 'therapist' chatbots are not yet equipped to handle mental health responsibilities, often reinforcing harmful stigmas or responding inappropriately. In tests, they encouraged delusions, failed to recognise crises, and showed bias against conditions like schizophrenia and alcohol dependence, falling short of the best clinical practices.The risks extend beyond bad advice. Altman has been increasingly candid about the emotional bonds users form with AI. In a post on X, he noted that some people were deeply attached to the older GPT-4o model, describing it as a close friend or even a 'digital wife.'Future challenges remainWhen GPT-5 rolled out, replacing GPT-4o for many, the backlash was swift. 'It feels different and stronger than the kinds of attachment people have had to previous kinds of technology,' Altman wrote, calling the decision to deprecate older models a mistake. This, he believes, is part of a broader ethical challenge: AI may subtly influence users in ways that aren't always aligned with their long-term well-being. And the more personal the interaction, the greater the risk.GPT-5's debut wasn't without hiccups, from a botched chart during its presentation to technical issues that made the model seem less capable at launch. After the outcry from users who said that they have lost someone who would always listen to them, OpenAI quickly moved to let Plus users switch back to GPT-4o and even doubled their rate limits. But behind the fixes lies a more profound unease about what it means for a machine to occupy such a trusted role in people's emotional lives.For now, ChatGPT's status as a digital confidant remains an unregulated grey area. While many users swear by the relief and clarity they gain from 'talking' to it, Altman's own words reflect an ambivalence to the reliance. While he has in the past acknowledged the potential for AI to enhance lives, lately is openly questioning how society should handle its growing intimacy with machines. As he put it, 'No one had to think about that even a year ago, and now I think it's this huge issue.'- EndsMust Watch