ChatGPT is today the go-to virtual confidant of millions of people around the world. Relationship problems to professional decisions, users are pouring their hearts out to the AI chatbot like it’s a trusted buddy ready to lend an ear at all times. But OpenAI head Sam Altman has a jarring revelation that will make you think twice before pouring your next personal crisis to the bot.
Talking on comedian Theo Von’s “This Past Weekend” podcast, Altman let loose a bomb most users probably never considered: your conversations with ChatGPT are not protected under any legal confidentiality laws. Unlike sitting with your therapist, your doctor, or your attorney, what you tell ChatGPT could find its way into a courtroom.
“These days, if you talk to a lawyer or a doctor or a therapist about all of that, there is privilege to talk to them legally. There is confidentiality. We haven’t worked that out yet with ChatGPT,” Altman said in the podcast.
This finding is particularly concerning due to the way the platform is being utilized. Young adults, in particular, have made ChatGPT their default counselor, presenting it with solutions for every minor annoyance of the day to major life decisions. The chatbot’s capacity to provide insightful, non-judgmental responses has given rise to a confidence that can be misplaced.
AI Privacy Perils and the Rise of Algorithmic Dependence
The absence of legal safeguards on AI chats is a massive risk. If your private interactions with ChatGPT find their way into a court hearing, OpenAI might be legally compelled to surrender them. Altman painted a disturbing scenario: “If somebody shares with ChatGPT their most private problems and that emerges in a courtroom, we might be legally required to provide that over. That is a real issue.”
The most impressive thing is how quickly this has arisen. Only a year ago, this privacy issue was not even on anyone’s radar. Now that AI-powered chatbots are becoming increasingly integrated into individuals’ lives, it’s quickly becoming an issue that regulators and tech firms have yet to address.
Aside from the issue of privacy, Altman has mentioned another disconcerting trend, which is among Gen Z users. In testifying before a Federal Reserve banking conference, he spoke of encountering young individuals who’ve become totally reliant on ChatGPT to make choices.
“There are teenagers who say, ‘I can’t make any choice in my life without asking ChatGPT everything that’s happening. It understands me. I will do whatever it tells me to do.’ That’s just terrible to me,” Altman said.
This level of dependence unsettles the CEO of OpenAI, even when the AI is providing good advice. He believes that there is something fundamentally incorrect about entrusting all thinking to a man-made system, however sophisticated.
Using ChatGPT for Mental Health, A Helpful Tool, Not a Human Replacement
While ChatGPT can provide useful advice and guidance, Altman reassures us that it is in no way a replacement for professional human care. A trained therapist has something AI does not: true human empathy, legal privilege, and decades of clinical education.
The bot may at times provide more acceptable advice from a human, but it does not have critical components such as context, real empathy, and responsibility. It also cannot provide the truly safe and trusting space needed for mental health treatment.
It does not mean you have to abandon ChatGPT altogether. The key is to use it responsibly. Use it as a tool to brainstorm or organize your thoughts, but not as your sole authority. When seeking advice, verify suggestions against human authority.
Most importantly, do not input sensitive, legal, or highly personal information that you would not want potentially disclosed.
For serious mental or emotional issues, expert therapists are still indispensable. They provide confidentiality, professional expertise, and human empathy that cannot be delivered by AI.
The more human and better AI gets, the more we risk forgetting that these are not human beings. Altman’s caution is a good reminder here that while ChatGPT can be a good tool, it can never replace human judgment, professional experience, or actual human interaction.
