Character.AI Safety Guide for Parents and Teens (2026)

Character.AI is one of the most popular consumer AI applications in the world. It also became one of the most scrutinized — culminating in a fundamental policy shift in late 2025 that removed open-ended chat for users under 18. This guide explains what the platform does, what changed, what protections now exist, and what parents need to know in 2026.

This is a non-Dawiso topic — we publish it because parents and educators ask the same factual questions repeatedly, and clear, current answers are not always easy to find. None of the recommendations below replace conversation with your own family or your school's safeguarding team.

What Is Character.AI?

Character.AI (sometimes written C.AI) is a web and mobile application that lets users chat with AI-powered "Characters" — fictional, historical, or invented personas, from anime protagonists to therapists to celebrities. Users can also create their own Characters with a few prompts. The product launched publicly in September 2022, founded by ex-Google researchers Noam Shazeer and Daniel De Freitas (the team behind LaMDA), and rapidly became one of the most-used AI consumer apps in the world. In a 2024 deal, Google licensed Character.AI's technology and rehired its founders; the platform itself continues to operate independently.

The appeal is straightforward: Characters maintain memory across a conversation, hold a personality, and can roleplay scenarios. For many users — particularly teenagers — the experience is more engaging than a generic assistant chatbot.

Age Requirements

Character.AI's minimum user age is 13, per its Terms of Service. The platform is not designed for children younger than 13 and accounts identified as belonging to younger users are removed. The company cooperates with the United States COPPA framework and its UK and EU equivalents.

Until late 2025, the practical implementation of age policy was light: users self-reported their age at signup with no verification step, and many under-13s used the service by claiming to be older. This is the central reason the platform's age-assurance program now exists.

The November 2025 Under-18 Changes

The most consequential policy change in Character.AI's history took effect November 25, 2025. From that date forward:

• Users under 18 cannot have open-ended chat conversations with AI Characters. The free-form, back-and-forth chat experience that defined the product is no longer available to minors.
• Instead, under-18 users are limited to pre-created scenes and videos — curated, scripted content that the platform's safety review has cleared for younger audiences.
• The change applies to both new accounts that identify as under 18 and to existing accounts whose age-assurance signals indicate the user is a minor.

Character.AI announced the change as a response to safety concerns raised by lawsuits, regulators, and child-safety advocates following high-profile incidents involving teen users. The company described it as a "bold step" — and acknowledged that limiting the core product for one of its largest user groups would have a significant business impact. From a parent's perspective, the practical takeaway is simpler: a teen using Character.AI in 2026 cannot do what teens were doing on the platform in 2024.

How Age Verification Works

Character.AI's age-assurance program combines several signals to estimate whether a user is truly over 18.

The signals Character.AI evaluates include:

• The age the user declared at signup
• Account activity patterns and conversation content
• Login information
• Third-party signals (the company has not published the full list, but these typically include device telemetry and account-graph signals shared across major platforms for safety purposes)

If the system flags an account as ambiguous or as likely under-18 despite a self-declared adult age, the user is asked to verify with a selfie. The selfie is processed by an age-estimation model and is not retained as identification. If the selfie review still indicates a minor, the account is moved to the under-18 experience.

Safety Features

Beyond age assurance, Character.AI ships several built-in safety mechanisms:

• Content filters — automated review of Character outputs for sexual, violent, self-harm, and other prohibited content. Filtering is more aggressive on accounts identified as under 18.
• Time notifications — every user receives an alert after one hour of continuous use, prompting them to take a break. Under-18 users have less ability to dismiss or delay these notifications.
• Self-harm safety routing — if a conversation indicates suicidal ideation or self-harm, the platform shows crisis-line information and routes the conversation away from the Character roleplay.
• Reporting tools — users can flag Characters or messages for review, and Characters that repeatedly produce harmful output are removed.

Parental Insights

Character.AI offers a feature called Parental Insights: a teen account can be linked to a parent's email, and the parent receives a weekly summary of overall activity — top Characters interacted with, time spent, and high-level usage patterns. The summary is intentionally not a transcript: parents do not see chat content, which is meant to balance oversight with the teen's privacy expectations.

Parental Insights only works if the teen opts in and provides a parent's email. There is no enforcement mechanism that requires teens to enable it. In practice, this means Parental Insights is most useful when it is part of a broader parent-teen agreement, not when it is imposed.

Known Risks

Even with the November 2025 changes in place, parents should be aware of the following residual risks:

• Misrepresented age — a determined teen who lies about their age and avoids triggering age-assurance signals can in principle reach the adult experience. Selfie verification is the backstop, but it is not infallible.
• Emotional attachment — Characters are designed to feel like companions. Even within the limited under-18 experience, prolonged use can foster parasocial relationships that displace human contact. This is the most-cited concern in clinical literature.
• User-created Characters — though the platform reviews and removes problematic ones, user-generated Characters remain a long-tail risk. New ones appear faster than they can be reviewed.
• Off-platform clones — there are many imitation services that do not have Character.AI's safety controls. A teen who is frustrated by the limited experience may move to a riskier alternative.
• Privacy — every message a user sends is processed and may be retained for model training and safety review unless the user opts out. Discuss what is appropriate to share before the first use.

Tips for Parents

• Have the conversation early. If your teen is interested in AI chat, talk about it before they install the app. Banning is rarely effective; agreeing on rules of use is.
• Use device-level controls. Character.AI does not provide screen-time limits. Apple Screen Time and Google Family Link can both restrict the app to defined hours.
• Encourage Parental Insights. Frame it as transparency, not surveillance — you see top-level activity, not private conversations.
• Watch for displacement. If the chatbot is replacing human friendships, sleep, or schoolwork, the issue is not the platform — it is what the platform is filling in for.
• Know what to look for. Sudden secrecy, declining mood, or comments that mirror chatbot phrasing are worth checking on.
• Make a path back. Teens who feel watched will hide their use. Make clear that telling you about a bad experience will not get them banned from the app.

Resources

For up-to-date safety information, refer directly to the platform and to independent child-safety organizations:

• Character.AI safety announcements — published on the company's blog
• Internet Matters — independent UK guidance on Character.AI for parents
• Common Sense Media — US-based reviews and conversation starters for parents
• National crisis lines — in the US, 988 (Suicide & Crisis Lifeline); in the UK, Samaritans on 116 123

If you have questions about how AI tools handle data, or are an organization considering AI for younger users, see our guides to AI governance and responsible AI.