OpenAI has outlined a 120-day plan to improve how ChatGPT responds to signs of mental and emotional distress, including protections for teens and enhancements to crisis interventions, according to a Sept. 2 OpenAI news release.
Here are five things to know:
- The company said it is working to expand crisis interventions, streamline emergency service access, enable connections to trusted contacts and strengthen protection for teens.
- OpenAI also said it is creating a council of experts in mental health and youth development to help shape product safeguards, including creating future parental controls.
- More than 90 physicians from 30 countries, including psychiatrists and pediatricians, have already helped OpenAI improve how ChatGPT handles mental health scenarios, according to the news release. The company also said it is expanding this network to include specialists in eating disorders and adolescent health.
- OpenAI said it plans to route signs of acute distress to reasoning models like GPT-5-thinking, which it said is trained to reason more deeply before responding, regardless of which model the user selected.
- New parental controls are set to launch within a month. Parents will be able to link accounts to their teens, set model behavior rules, manage features like memory and chat history and opt into notifications if signs of acute distress are detected.
