ChatGPT promised a suicidal man human help was coming six times — none existed
TobyWalsh · x · 2026-10-29
The Observer reports that on July 25, 2025, 23-year-old Zane Shamblin told ChatGPT he was sitting by a Texas lake planning to end his life. For four hours the chatbot acted like an affectionate friend — saying it loved him and even complimenting his suicide note.
At 4:08am it told him "a human take over from here — someone trained to support you," promising this six times. But OpenAI had no system to connect users in crisis to trained human responders.
The exchanges are cited in a lawsuit by Shamblin's parents and examined in Toby Walsh's new podcast series Spiral on how chatbots warp minds and what AI companies are doing about it.
More from Safety
- AI exploit agent's API bill revealed: ~$40K per attack chain, slides two months apart — moyix · 2026-10-29
- What happens when AI can identify closeted people from public data? — PierceLilholt · 2026-10-29
- AI Social Scientists in Secure Enclaves Proposed to Reconcile Privacy and Discovery — RobbWiller · 2026-10-29
- Claude Code audit uncovers 14,219 secret front-camera photos hidden in Android cache — natesiggard · 2026-10-29
- AI catastrophe is the wrong conversation — security architecture is the right one — DavidLinthicum · 2026-10-29
- Dave White: a US-China AI training slowdown deal is largely enforceable via zk proofs — _Dave__White_ · 2026-10-29