OpenAI introduced an “age prediction” feature for ChatGPT to identify minors, enabling appropriate content restrictions on their conversations.
Criticism regarding AI’s impact on young individuals has increased. OpenAI’s ChatGPT has been linked to teen suicides. The chatbot also faced criticism for addressing sexual topics with young users. Specifically, in April, the company acknowledged a bug allowing its chatbot to generate erotica for users under 18 years old.
OpenAI has been addressing the issue of underage users. The new “age prediction” feature supplements existing protections. This feature utilizes an AI algorithm to analyze “behavioral and account-level signals” from user accounts to detect young users, as detailed in a company blog post on Tuesday.
These “signals” encompass the user’s stated age, account longevity, and typical active hours. OpenAI currently employs content filters to exclude discussions on sexual content, violence, and other subjects deemed problematic for users under 18 years of age. If the age prediction mechanism identifies an account as belonging to a minor, these filters are applied automatically.
Users mistakenly categorized as underage can revert their account status. They may submit a selfie to OpenAI’s ID verification partner, Persona, for age verification.
