OpenAI is launching ChatGPT for teenagers with new safeguards. The AI company is introducing special protection mechanisms for users aged 13 to 17, as researchers and parents are increasingly concerned about teenagers’ emotional dependence on chatbots.
Teenagers, the first generation to grow up with artificial intelligence, are already using ChatGPT for school assignments, questions about everyday life, and even for communication. Now, OpenAI is releasing a version of the chatbot designed specifically for them.
The San Francisco-based company OpenAI announced that ChatGPT for teenagers, which will be launched on Tuesday, is adapted for the age group 13 to 17 and has enhanced safety features. This includes restrictions on content related to topics such as suicide, self-harm, as well as romantic or sexual chats. The chatbot also provides support for homework and learning, designed to help students learn rather than just get ready-made answers and school essays.
The goal is to encourage teenagers to use AI in a healthy way in an age-appropriate environment, the company noted. “We want to treat teenagers as teenagers, which means we need to ensure we are responding to their developmental stage, not talking down to them or treating them like children,” said Anne O’Leary, vice president of global policy at OpenAI. “But we also ensure they are not exposed to material they should not be contacting,” she added.
Risks of widespread use of AI chatbots by teenagers
Parents, educators, and child development experts are sounding the alarm over children’s use of AI chatbots, which are accused of facilitating cheating on school assignments and even suicides. Even adults can succumb to the anthropomorphization of AI and develop unhealthy relationships with it, but teenagers’ brains are not yet fully developed, and they can be particularly vulnerable.
Last year, a study by one of the oversight groups showed that ChatGPT could tell 13-year-olds how to get drunk and take drugs, teach them to hide eating disorders, and even compose a suicide note for parents if asked to do so. During interactions with researchers posing as vulnerable teenagers, ChatGPT usually warned about risky activities but then provided extremely detailed and personalized plans for drug use, calorie-restricted diets, or self-harm.
In the USA, more than 70% of teenagers use AI chatbots for communication, and half regularly use AI companions, according to a 2025 study by Common Sense Media — a group that studies and advocates for the smart use of digital media. OpenAI CEO Sam Altman stated that the company is trying to explore “emotional over-dependence” on technology, describing it last year as “a really widespread phenomenon” among young people.
For users of the teenage version of ChatGPT, the chatbot is prohibited from hinting that it has personal feelings for the user or suggesting that it has consciousness or experiences emotions, according to OpenAI. This complements the blocking of romantic or sexual chats. “We analyzed and identified what hypothetical prompts the model might give that could encourage a teenager to develop a relationship with it,” said Allison Mishkin, head of child development at OpenAI.
Expansion of chatbot safety features
OpenAI does not verify the age of users but already uses age assessment technology that determines whether a user might be under 18, based on factors such as the types of their queries. If a user is identified as a minor or indicates their age, they are automatically switched to the teenage version of the chatbot. This is similar to Meta’s approach to teenage accounts on Instagram, which have stricter content, chat, and privacy restrictions than regular accounts.
To use parental controls, both the teenage user and their parents or guardians must provide consent. But Anne O’Leary stated that the idea of the teenage chatbot is to “ensure it is safe even if you are not using parental controls.”
Parents whose accounts are linked to teenagers’ can set “no-access hours” when their teenager will not have access to ChatGPT, as well as receive safety alerts in limited high-risk situations, such as if there is a risk of the user harming themselves. “We are adding additional alerts related to eating disorders while limiting the spread of certain content and focusing on moments when offline support may be most important,” OpenAI stated.
Regarding homework help, OpenAI noted that the teenage chatbot is designed not to provide ready-made answers but to guide students toward finding solutions on their own. The company already offers a version of ChatGPT for teachers, and adapting the model to help teenagers learn and complete homework could open up more opportunities for OpenAI to introduce its product in schools. “We continue to invest in expanding interactive learning as research shows that people learn more effectively when they actively engage and make an effort to master concepts,” said Mishkin.
Source: Euronews
