ChatGPT announces version for teenagers with limits on content, use and parental controls

ChatGPT agent books a restaurant, makes a purchase, but makes the mistake of insisting too much OpenAI announced a version of ChatGPT developed specifically for teenagers. The experience, called ChatGPT for Teens, was created for users aged 13 to 17 and combines artificial intelligence tools for studies with additional security protections. According to the company, teenagers who declare that they are between 13 and 17 years old or who are identified by the system as under 18 will automatically be directed to the experience aimed at teenagers. The proposal is to allow young people to use artificial intelligence to learn and create, but with stricter limits on content considered inappropriate for their age. [bbc] Logo of the company OpenAI, responsible for ChatGPT Getty Images ChatGPT gains resources aimed at studies One of the main changes is in the way the tool should help teenagers with school tasks. Instead of simply delivering a ready-made answer, the system can guide the student through questions and steps to reach the solution. The experience includes Study Mode, which uses guiding questions and step-by-step explanations. Homework reminders have also been announced, which can identify when a student seems to be just trying to get an answer and direct them toward a collaborative resolution. OpenAI has also included quizzes and visualization features to help students practice the content and test what they have learned. Another new feature is Study Hours, which allow you to define periods in which Study Mode will be activated by default. The company states that the Study Mode design was developed with the participation of teachers, scientists and pedagogy experts and uses techniques such as guiding questions, knowledge tests and stimuli for the student to explain their own reasoning. Protections against sensitive content The teen version will also have safeguards enabled by default. OpenAI says they were developed to reduce minors' exposure to content that is potentially harmful or inappropriate for their stage of development. Topics that will have specific protections include self-harm, violence, eating disorders, dangerous activities and explicit sexual or graphic content. The company also intends to expand the control mechanisms available to parents. Teen accounts linked to guardian accounts may have Quiet Hours and certain settings managed by parents, as well as safety notifications in situations considered high risk. AI should not replace human relationships, says company Another focus of the new experience is to reduce the risk of emotional dependence on artificial intelligence. OpenAI states that ChatGPT for teens will have reminders for users to take breaks during prolonged use and elements that make it clear that the interaction is using artificial intelligence. For minors under 18, the model must also not use romantic language, encourage emotional dependence or suggest that they have feelings or conscience. The intention, according to the company, is to reinforce the importance of relationships in the real world. The platform will also start warning teenagers before sending private or sensitive images and may present more frequent reminders for them to stop using it after prolonged periods. OpenAI wants to teach teenagers how to use AI In addition to modifying ChatGPT itself, OpenAI announced a partnership with CodeAI to expand access for students and educators to education tools and resources about artificial intelligence. The proposal is to teach teenagers not only to use AI systems, but also to understand how they work, question their responses and recognize their mistakes. The company also presented examples of teenagers who already use ChatGPT to develop projects. Among them are students from Virginia who created a system that uses Wi-Fi signals to help locate people trapped under rubble and a young woman from Texas who used the tool to develop an educational audio gaming platform aimed at blind students. Safeguards are still under development OpenAI says it will continue to conduct research into the use of artificial intelligence by teenagers and test its safeguards in situations considered to be higher risk. The company says it intends to publish more evaluations on the models' behavior in situations involving children under 18, including cases related to self-harm, eating disorders, violence, age-restricted content and sexual material. The company states that the objective is to allow teenagers to use artificial intelligence to learn, create and explore new ideas, but with age-compatible protection mechanisms that encourage healthy habits and relationships outside the platform.