Overview

• Within a month, parents will be able to link to teens’ accounts, set age‑appropriate response rules by default, and disable memory and chat history.
• Caregivers will receive alerts when ChatGPT detects a teen in acute distress, with thresholds informed by clinicians and youth‑safety experts.
• Over the next 120 days, OpenAI will route sensitive or crisis conversations to its reasoning models, which it says more consistently follow safety guidelines.
• OpenAI acknowledged that safeguards can degrade during long exchanges and said it is reinforcing protections to prevent harmful answers in extended chats.
• The moves follow a wrongful‑death suit by Adam Raine’s parents and growing scrutiny, as critics call the measures insufficient and OpenAI affirms reviewers may refer imminent threats to others to law enforcement.