AI boss warning: How easily the state can see what you write to ChatGPT
Published 12 September 2025 at 13.25

Foreign. Conversations with AI should be protected in the same way as letters or meetings with doctors and lawyers. This is according to Sam Altman, CEO of OpenAI, who warns in an interview that the state today can request users' chats with ChatGPT far too easily.

"We should have something similar to physician or lawyer's secrecy even for AI," Sam Altman said in an interview with Tucker Carlson.

“When you tell us about your health or legal problems for ChatGPT, the government should not be able to access it. Right now, they can actually do it.

Altman said he himself had been in Washington arguing for a new piece of legislation that would protect people's AI calls.

According to its own policy, OpenAI does not sell data, but that current laws still allow authorities to access information via legal processes relatively easily.

After warning of transparency in people's most private issues during the interview, Altman also received questions about AI's impact on the labour market.

He specifically singled out customer service jobs as an area that is likely to disappear at a rapid pace.

"Many of the simpler customer service roles at computer or phone will simply be replaced by AI," he said.

“Glorrors will also be heavily affected, but there it may as well be a redesign of the job as it disappears.

At the same time, Altman argued that professions that require human presence and empathy, such as nurses, are less threatened:

“When you’re sick, you want to meet a person, not just a machine.