OpenAI Introduces ‘ChatGPT for Teens’ as Safety Concerns Grow

OpenAI announced on Tuesday “ChatGPT for Teens,” a new mode in which the chatbot automatically limits some conversations to better protect young users, as scrutiny grows over the harmful effects of artificial intelligence.

Many features of the new mode deal with safety, the company said in a blog post. The chatbot will limit high-risk conversations that involve self-harm, violence, eating disorders, and *** ***ual and graphic content. In some cases, ChatGPT will alert adults who have opted in to parental controls and notifications. And the mode will strengthen protections that prevent the chatbot from suggesting it has personal feelings.

ChatGPT’s teen mode also contains new studying tools and guardrails for students, OpenAI said. If the chatbot detects that a teenager is working on homework, it should respond to prompts with further questions instead of just providing the answer, for example.

OpenAI requires people to provide their age while signing up for ChatGPT — under the terms of use, they must be 13 or older — but some lie. The company tracks more than 2,000 signals to detect whether a user is under 18, it said, in which case the chatbot will automatically turn on the teen mode.

A company spokesman said the signals, which can include typical login times and the length of the account’s existence, were “minimally privacy invasive.”

OpenAI, which requires parental consent for teenagers to use its chatbot, has faced mounting pressure over the safety, mental health and development of its younger users. A growing number of families have sued the company, claiming their children were exposed to dangerous content from the chatbot and engaged in self-harm or died by suicide as a result.

In June, Florida became the first state to sue OpenAI, accusing the company of misrepresenting the safety of its platform for minors.

Parents’ groups in New York City, Los Angeles and elsewhere are pressing schools to temporarily halt rollouts of A.I. products for students, concerned that chatbots could hinder or diminish learning. Chatbots regularly gin up misinformation and can help students cheat. Some education researchers warn that the A.I. tools can impede young people’s critical thinking.

OpenAI said its new study mode was an additional educational tool for teenagers.

“We believe AI should expand educational opportunity, which is why we designed ChatGPT for Teens to support moments outside the cl***room when help is not always available, while keeping teens engaged in an active and collaborative learning process,” the company said in its blog post.

OpenAI’s approach to tweaking ChatGPT for teenagers follows a similar playbook used by social media companies for years, in which new youth modes with guardrails are created but use of the app is still encouraged. The new features also typically rely on parental controls and monitoring, something that child-safety advocates and lawmakers have criticized as displacing the responsibility for harms ultimately created by the companies.

Additionally, verifying the age of users has been a problem for tech companies because of easy workarounds.

(The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news content related to A.I. systems. The two companies have denied those claims.)

Natasha Singer contributed reporting.