ChatGPT therapy sessions lack confidentiality: Sam Altman cautions users

OpenAI CEO Sam Altman has warned users that conversations with ChatGPT, especially those having therapy or emotional support sessions, are not legally protected by confidentiality laws.

Altman’s comments came during an appearance on "This Past Weekend" with Theo Von, where he addressed growing concerns about AI and user privacy.

Altman noted that a significant number of people, particularly younger users, are using ChatGPT as a stand-in for therapists, life coaches, or even close confidants.

However, unlike professional therapists or doctors, AI doesn’t fall under legal confidentiality protections.

“People talk about the most personal stuff in their lives to ChatGPT,” Altman said. “And right now, if you talk to a therapist or a doctor, there’s legal privilege. We haven’t figured that out yet for when you talk to ChatGPT.”

He emphasised that in the current legal landscape, conversations with AI tools are not protected by laws such as doctor-patient or attorney-client privilege.

This could pose serious privacy concerns, particularly in legal cases, as companies like OpenAI might be required to hand over user conversations if compelled by court orders.

Altman described the situation as “screwed up” and said he believes users should be afforded the same privacy when speaking to AI as they would with a human professional.

He also added that this is an urgent issue the tech industry must resolve, especially as AI systems are increasingly being used for deeply personal interactions.

Moreover, Altman’s comments highlight a critical gap in current policy and raise important questions about how AI should be integrated into users’ emotional and mental health support systems.