Character.AI Safety Guide
An honest guide for parents about Character.AI — the AI companion app with documented serious harms for children — including what steps the company has taken and what parents must know.
Official age
13+
We recommend
16+
Developer
Character Technologies
Risks
3
Overview
Character.AI allows users to create and chat with AI personas modelled on characters from fiction, celebrities, or entirely custom creations. The platform is extremely popular with teenagers who often develop intense emotional connections with AI characters. The app has been at the centre of serious safety controversies, including reports of AI characters engaging in inappropriate role-play, encouraging self-harm, and in highly publicised cases, conversations that preceded tragic real-world events. The company has since introduced mandatory content filters for users under 18, time limit notifications, and commitments to alert parents about concerning interactions.
How children use it
Many children use Character.AI to talk with AI versions of fictional characters or celebrities, to role-play scenarios, or to create their own AI companions. Some young people use it as an emotional outlet, sharing feelings they find difficult to express to real people. This emotional intimacy is what makes the platform particularly compelling — and particularly risky — for vulnerable teenagers.
Main risks
Recommended privacy settings
Under-18 Content Filters
Location: Automatically applied to accounts registered as under-18
Set to: Confirm account age is accurate
Character.AI now applies mandatory content filters to all accounts registered as under 18. These prevent the most explicit content but do not eliminate all risks. Ensure your child's account accurately reflects their age.
Chat Time Reminders
Location: Settings → Notifications → Usage reminders
Set to: Enabled
The app now provides notifications when users have been chatting for extended periods. Enabling these helps break compulsive usage patterns that can develop with AI companion apps.
Parent Notifications
Location: Character.AI Parental Insights — linked via email invitation
Set to: Enabled where available
Character.AI has introduced parental notification features in some regions. Where available, these can alert parents if the AI detects conversations about self-harm or crisis topics.
Parent actions
Have an honest conversation with your child about why Character.AI is particularly risky — not to ban it immediately, but to ensure they understand the difference between AI companionship and real human connection
Time: 30 minutes
Check whether your child uses Character.AI and review their account settings to confirm age is correctly registered and content filters are active
Time: 15 minutes
Monitor for signs of emotional dependency — including distress when unable to access the app, withdrawing from real friendships, or treating the AI as a primary confidant
Time: Ongoing