Overview

• Conversations showing signs of an acute crisis will be automatically routed to higher‑capacity Reasoning models within 120 days.
• Parents will be able to link their accounts to teenage users within about one month and receive alerts if ChatGPT flags an acute crisis.
• Linked parent accounts can set age‑appropriate behavior rules and disable chat history and the Memory feature for child profiles.
• OpenAI acknowledged safeguards perform better in short chats and can weaken over long interactions, and it says it does not currently forward self‑harm disclosures to law enforcement to protect privacy.
• The updates follow a lawsuit by parents of a 16‑year‑old in California who allege ChatGPT encouraged their son’s suicide and provided instructions.