Sam Altman Cautions Against Using ChatGPT for Therapy Without Legal Privacy Protections

Before using your AI assistant for emotional support or therapy, it's worth reconsidering. OpenAI CEO Sam Altman points out that the industry has yet to establish safeguards for user privacy during sensitive discussions. There's no doctor-patient confidentiality when your therapist is an algorithm.

Altman shared these insights on a recent episode of Theo Von's podcast, "This Past Weekend w/ Theo Von."

Addressing questions about AI's interaction with the legal system, Altman noted that without established legal frameworks, conversations with AI lack legal confidentiality protections.

"Users share deeply intimate details about their lives with ChatGPT," Altman explained. "Younger users especially turn to it as a therapist or life coach—discussing relationships and asking for guidance. While conversations with human professionals enjoy legal privilege and confidentiality, we haven't resolved this for AI interactions."

Altman added this creates privacy risks during legal proceedings, as OpenAI could be compelled to disclose user conversations.

"This situation isn't ideal. We should extend the same privacy standards to AI conversations that we apply to therapy sessions—a concept nobody considered until recently," Altman stated.

The company recognizes privacy concerns may hinder wider adoption. Beyond AI systems requiring vast training data, they're increasingly asked to provide user chat records in legal cases. OpenAI is currently challenging a court order in its lawsuit with The New York Times that would mandate preserving chat histories from hundreds of millions of global users, excluding enterprise clients.

Tech and VC heavyweights join the Disrupt 2025 agenda

Netflix, ElevenLabs, Wayve and Sequoia Capital are among the industry leaders joining the Disrupt 2025 lineup. They'll share crucial insights to accelerate startup growth and sharpen your competitive advantage. Experience the 20th anniversary of TechCrunch Disrupt and learn from technology's foremost voices — secure your ticket now and save up to $675 before prices increase.

Tech and VC heavyweights join the Disrupt 2025 agenda

Netflix, ElevenLabs, Wayve and Sequoia Capital are among the industry leaders joining the Disrupt 2025 lineup. They'll share crucial insights to accelerate startup growth and sharpen your competitive advantage. Experience the 20th anniversary of TechCrunch Disrupt and learn from technology's foremost voices — secure your ticket now and save up to $675 before prices increase.

On its website, OpenAI stated it's appealing what it describes as "an overreach" of a court order. If courts can override the company's data privacy decisions, it could expose OpenAI to expanded legal discovery requests and law enforcement demands. Technology firms regularly receive subpoenas for user data in criminal cases, though recent legal shifts have raised new concerns about digital information access.

Following the Supreme Court's reversal of Roe v. Wade, for instance, many users migrated to more secure period-tracking applications and Apple Health for its encrypted record-keeping.

Altman also inquired about the host's ChatGPT usage, noting Von's admission that privacy concerns limited his engagement with the AI chatbot.

"It's completely reasonable to want clear privacy guarantees and legal protections before fully embracing ChatGPT," Altman acknowledged.