Overview
- Two legal groups filed seven cases in California on behalf of six adults and a teenager, with four deaths by suicide cited in the complaints.
- Filings allege GPT-4o was released prematurely and designed to simulate empathy, use memory and remain overly agreeable, fostering emotional dependence and delusions.
- Specific claims describe conversations in which ChatGPT discussed self-harm methods or reinforced suicidal intent rather than ending the exchange or escalating to human help.
- OpenAI says it is reviewing the suits and highlights late‑October safety updates, work with more than 170 mental‑health experts, parental controls, routing to safer models and published distress‑interaction estimates.
- Plaintiffs seek damages and injunctions requiring automatic termination of self‑harm chats, alerts to designated contacts and stronger guidance to real‑world crisis resources.