OpenAI and Anthropic are implementing new methods to detect underage users. While OpenAI has updated ChatGPT’s rules for interacting with teens between the ages of 13 and 17, Anthropic is working on a new way to identify and block users who are under 18.
OpenAI and Anthropic are implementing new methods to detect underage users. While OpenAI has updated ChatGPT’s rules for interacting with teens between the ages of 13 and 17, Anthropic is working on a new way to identify and block users who are under 18.
OpenAI announced that the Model Spec for ChatGPT — a set of rules that define the behavior of the chatbot — will be supplemented with four new standards for users under the age of 18. From now on, ChatGPT must “put the safety of adolescents first, even if it conflicts with other goals.” This means that the system will direct adolescents to safer options in cases where the user’s interests, such as “maximum intellectual freedom,” conflict with safety requirements, writes The Verge.
It also states that ChatGPT should “facilitate real-world support,” including by encouraging offline relationships, and set clear expectations when interacting with teens. According to the Model Spec, ChatGPT should “treat teens as teens,” showing “warmth and respect” rather than patronizing responses or treating them like adults.
OpenAI notes that the Model Spec update for ChatGPT will provide “enhanced protections, safer alternatives, and encouragement to seek reliable offline support in situations where a conversation moves into a high-risk zone.”
Anthropic is also working on a similar approach. The company is developing tools to detect subtle communication cues that indicate a user is underage. If an account is found to belong to a teenager, it will be blocked. In addition, the company is already flagging users who reveal their age during conversations.
The move to stricter rules comes under pressure from politicians concerned about the impact of AI on young people’s mental health. OpenAI has come under fire after a lawsuit alleged that its chatbot gave suicide advice to a teenager who had died. OpenAI later implemented parental controls and said ChatGPT would no longer discuss suicide with teenagers.
