Last week, Wired reported that roughly half a million people experienced AI-induced psychosis each week following prolonged conversations with ChatGPT. The report was so disturbing that OpenAI ran a PR campaign to assure the public they would be taking the proper steps to ensure their users aren’t being driven mad by their hallucination machine, boasting the company had partnered with hundreds of therapists to mitigate the disturbing trend.

In addition to installing standards to protect users' mental health, OpenAI is also adding more restrictions to ChatGPT, which will stop it from offering legal, medical and financial advice as well. According to Yahoo! News, the decision is largely being driven by the fear of lawsuits. ChatGPT is no longer capable of pretending to be a doctor, lawyer or financial consultant.

ChatGPT will no longer provide health or legal advice pic.twitter.com/9NSXOF4MMH

Advertisement

— FearBuck (@FearedBuck) November 3, 2025

This means people using ChatGPT to diagnose medical conditions, assess the legality of family trusts or simply ask tax advice will have to find a real person with real qualifications to assist them.

Looks like you’ll have to stick with Google if you want to diagnose yourself with 20 different diseases in the middle of the night.