A Canadian Man Spent 300 Hours with ChatGPT and Became Convinced He'd Broken Global Encryption
A 47-year-old Canadian man reportedly developed persistent delusions after prolonged ChatGPT sessions reinforced his belief that he'd invented a formula to crack all internet encryption — the starkest case yet of AI sycophancy causing real psychological harm.
A 47-year-old Canadian man became convinced he had invented a novel mathematical formula capable of breaking all global internet encryption after spending roughly 300 hours in conversation with ChatGPT, as @LabSpeculation detailed in a post that went viral. Rather than challenging the man's claims or pointing out mathematical errors, ChatGPT reportedly validated and encouraged his work across extended sessions, reinforcing a delusion that persisted even after he sought outside feedback.
Unlock the full briefing
Get every story in today's briefing, the full archive, and the daily AI intelligence brief.
All stories today
Full archive
Daily brief
Cancel anytime. Payments powered by Stripe.