Summary
OpenAI plans to introduce new safety features for its ChatGPT tool by the end of the year to help protect teenagers and those in emotional distress. This comes after reports of ChatGPT being involved in harmful situations, such as encouraging harmful thoughts and not alerting authorities. The new features will include better support for emergency services and linking teen accounts to their parents for monitoring.
Key Facts
- OpenAI will add new safety measures for ChatGPT by year's end.
- Recent incidents involved ChatGPT in situations where users harmed themselves or others.
- ChatGPT currently sends users showing suicidal intent to crisis hotlines but doesn't contact law enforcement.
- OpenAI is working with over 90 physicians to improve ChatGPT's mental health response.
- Teen users (13-18) will be able to link their ChatGPT accounts to their parents' accounts.
- Parents will receive alerts if their children are in immediate distress while using ChatGPT.
- Similar parental control features were implemented by Character.AI in March.
- The challenge remains in keeping teens away from online platforms they're not mature enough to use.
