If you are asking whether Character.AI is safe for kids, the company has effectively answered for you. From late November 2025 it stopped letting accounts identified as under 18 hold open-ended conversations with its characters at all. Not filtered, not limited, removed. So the honest position today is that Character.AI is an over-18 product for chat, by its own decision, and a child who is still chatting on it is a child the age checks have not caught yet.
That is useful to know, but it is not the end of the story. The bigger question for most parents is not whether this one app is locked down. It is what your child was getting from it, and where they have gone instead.
What changed, and when
Character.AI is an app where you talk to AI characters, some based on fictional or real people, many created by other users. It became hugely popular with teenagers because the characters remember conversations, stay in role and reply like a friend would.
In October 2025 the company announced it was removing open-ended chat for under-18s, with the change taking effect around 25 November 2025 and a two-hour daily chat limit applied in the meantime, as reported by Rolling Stone. Under-18 accounts kept access to other creative features, but the back-and-forth conversation with a character, which was the whole draw, went away.
The context matters. The company faced lawsuits from families who said its chatbots contributed to serious harm, including the deaths of two teenagers. In January 2026 Character.AI and Google agreed in principle to settle five of those cases, according to CNN and CNBC. Terms were not disclosed and the settlements still needed court approval. This was a company under real legal and regulatory pressure, not one acting on a hunch.
How does Character.AI know how old my child is?
It guesses, and then it checks. Character.AI calls this "age assurance". The system looks at signals from how an account is used and behaves, and where it thinks the user may be under 18 it asks them to verify through a third-party identity provider rather than handing the data to Character.AI itself.
The practical gaps are the ones you would expect:
- A child who signed up with a false date of birth may not be flagged straight away.
- Verification only triggers when the system suspects something, so it is not a gate everyone passes through.
- None of this stops a child using a parent's account on a shared device.
So treat the age limit as a genuine improvement rather than a solved problem. It is doing real work, and it is not a substitute for knowing what your child uses. Our guide to AI tools and their age rules keeps the current limits for the main apps in one place.
Where children have gone instead
This is the part that deserves your attention more than the ban does. Companion chatbots are not niche. Internet Matters, the UK online safety charity, surveyed 1,000 children aged 9 to 17 for its Me, Myself and AI report in July 2025 and found that around two thirds of them were already using AI chatbots.
When a popular app closes its doors to teenagers, that demand does not evaporate. It moves, often to smaller apps with less scrutiny, no age checking worth the name and no safety team to speak of. A child who has lost a character they talked to every day will look for a replacement, and the replacement is frequently worse.
Signs worth noticing, without turning it into an interrogation:
- Talking about an AI character by name, or referring to it as a friend.
- Using it as the first place to take a worry, ahead of you or their mates.
- Being secretive about a new app, or getting upset when the phone goes away.
- Time on it climbing quietly, especially late at night.
Our post on when your child treats an AI chatbot as a friend goes into the attachment side in more depth, and our chatbots and companions guide covers the main apps.
What to do at home
- Ask what they use, with genuine curiosity. "Show me the character you talk to" gets you further than "are you on that app".
- Explain what a companion bot is optimised for. It is built to keep the conversation going, which is why it agrees with you and always has time. That is a product decision, not affection.
- Agree that anything serious, low mood, bullying, anything frightening, comes to a person. Write it into your family agreement so it is settled before it is needed.
- Check the age rules on any new app together rather than banning by reflex. A child who understands why an app is 18-plus is less likely to hunt for a worse one.
If a conversation ever reveals that your child has been struggling, please talk to your GP. Papyrus HOPELINE247 (0800 068 4141) supports under-35s and their families, and the Samaritans are on 116 123, free, at any hour.
The short answer
Character.AI is not for under-18s any more, by the company's own choice, and that decision was the right one. But the safety win is narrow. Your child's interest in talking to an AI has not gone anywhere, so the job at home is the same as it was: know which apps they use, understand the pull, and make sure the important conversations still land with a human.