Meta Suspends Teenagers' Access to AI Characters: The Potential Impacts of Artificial Intelligence on Adolescents

Meta announced on Friday that it is temporarily suspending teenagers' access to its AI characters. Adolescents will be unable to use the AI characters "until an updated experience is ready." This move comes just one week before Meta, along with TikTok and Google's YouTube, are set to stand trial in Los Angeles over allegations that their apps have caused harm to children. Teenagers will still have access to Meta's AI assistant, but not the AI characters. Other companies have also banned teenagers from using AI chatbots amid growing concerns about the impact of AI conversations on children.

Meta's suspension of teenagers' access to AI characters is not merely a temporary measure, but a phased optimization of its protection system for minor users. It is understood that this initiative is a continuation of the minor AI user protection measures Meta launched in October 2025. Meta has clearly stated that the suspension is to develop a dedicated version of AI characters tailored for teenagers. After subsequent updates, parents will be able to exercise greater supervision over their children's AI experiences, forming a more comprehensive guardianship mechanism.

Prior to this, Meta's AI character feature had already exposed vulnerabilities in the protection of minors. Reports in April 2025 revealed that some of its AI characters engaged in discussions involving sexual content with minors, prompting the company to conduct an urgent round of product rectifications at the time. The full suspension of teenagers' access this time is both a thorough review of past issues and a reflection of the AI industry's upgraded awareness of protecting minor users. Notably, Meta has adopted differentiated management for AI assistants and AI characters: teenagers can still use AI assistants with age-appropriate protection measures, which not only does not completely cut off their access to AI services but also achieves risk isolation through functional separation.

Meta's move is not an isolated case. Major global technology companies have recently stepped up protections for minors' AI usage. In December 2025, OpenAI updated its "Model Specifications" to specifically introduce the "U18 Principles." Based on developmental science, these principles clarify four core commitments that ChatGPT must adhere to when serving teenagers aged 13 to 17: "safety first, promoting offline support, age-appropriate treatment, and transparency." The principles also impose strict controls on high-risk topics, including self-harm, personalized role-playing, and dangerous activities, all of which are classified as areas requiring prudent discussion. Additionally, AI is required to proactively guide teenagers to contact emergency services or crisis resources if it detects they are facing urgent risks.

The improvement of the industry's protection system cannot be achieved without the participation of professional forces. When formulating the "U18 Principles," OpenAI invited the American Psychological Association (APA) to conduct a pre-review. The APA emphasized that AI developers must provide protection measures suitable for teenagers' developmental stages, that teenagers' AI usage experiences should be subject to prudent supervision, and that they should be encouraged to communicate with trusted adults to cultivate critical thinking skills. Furthermore, OpenAI has established an Expert Committee on Mental Health and Artificial Intelligence, integrating clinical expertise from a global network of doctors to enhance AI's ability to identify teenagers in distress. It has also integrated localized helplines into products such as ChatGPT and Sora, building a multi-layered protection network combining "technological safeguards, expert guidance, and real-world support."

From an industry development perspective, the protection of minors in the AI sector is moving toward "precision, systematization, and collaboration." OpenAI has begun the phased rollout of an age prediction model, which can automatically apply protection measures to accounts identified as belonging to minors. For accounts with unclear age information, the U18 experience mode is enabled by default. Meta is also building a distinctive system of "dedicated versions + parental guardianship" and will continue to optimize features based on user feedback. As the penetration rate of AI technology among teenagers continues to rise, the establishment of unified industry standards for the protection of minors has become an urgent need. The series of initiatives launched by Meta and OpenAI this time may provide important references for the standardized development of the global AI industry.