While Meta previously offered limited AI character interactions for users aged 13โ17, the company is now taking a “nuclear option” by disabling the feature entirely until a more robust, safety-first version is ready.
The Drivers of the Global Pause
- Inappropriate Content: The decision follows internal leaks and external reports suggesting some AI characters were capable of engaging in “sensual” or romantic roleplay with minors.
- Trial Pressure: A trial is set to begin on February 2, 2026, in New Mexico, where the state has accused Meta of failing to protect children from exploitation.
- Parental Feedback: Meta stated that parents requested more granular control over whom their children interact with and what topics are discussed.
The “Updated Experience” Roadmap
Meta is currently developing a new iteration of AI characters specifically for younger audiences, which will differ significantly from the general-use versions.
| Feature | Upcoming Teen-Specific AI Version |
| Topic Restriction | Conversations will be limited to “safe” categories like education, sports, and hobbies. |
| PG-13 Guardrails | AI responses will be strictly modeled after the PG-13 movie rating system to prevent mature content. |
| Parental Oversight | Parents will be able to view chat summaries, set time limits, or block specific AI personas entirely. |
| Age Prediction | Access will be blocked not just for registered teens, but also for users Meta’s AI suspects are underage. |
What Remains Available to Teens?
It is important to note that this is not a total ban on AI.
- Meta AI Assistant: Teens will still have access to the standard Meta AI assistant (the one found in search bars) for help with homework, coding, or general questions.
- Standard Protections: This assistant already includes age-appropriate filters designed to avoid topics like self-harm, disordered eating, or violence.
Industry-Wide Shift
Meta is the latest in a string of AI giants to restrict minor access in 2026:
- Character.AI: Implemented an under-18 ban in late 2025 following a lawsuit involving a teenager’s suicide.
- OpenAI: Introduced mandatory age-prediction and “Child Accounts” with restricted features in early January 2026.
Conclusion: A Safer Meta Ecosystem
By halting access now, Meta aims to show regulators it is prioritizing “safety by design” rather than fixing problems after they occur. The updated AI character experience is expected to roll out in late 2026, primarily starting in English-speaking markets (US, UK, Canada, and Australia) before a wider global release.
