• What is Character.AI
• Why your child might be using Character.AI
• Official Character.AI age restrictions and how they’re enforced
• Is Character.AI included in the social media ban?
• What Character.AI costs
• Privacy: what happens to your child’s data when they use Character.AI
• How to set up Character.AI parental controls
• What to watch out for
• Fiona’s take on Character.AI

---

What is Character.AI?

Character.AI is a chatbot app where people create and talk to AI “characters” built around a personality. A character might be invented, based on a fictional character, or modelled on a type of person, such as a therapist or tutor. This is often referred to as an AI companion.

Why your child might be using Character.AI

Children and teenagers use it to roleplay, write stories with an AI, or talk to a character for company. It’s built around ongoing conversation with a character, not quick factual answers.

Official Character.AI age restrictions and how they’re enforced

Character.AI’s own terms set the minimum age at 13. But on 25th November 2025, the company went further and now nobody under 18 can have an open-ended conversation with a character, anywhere the app operates, including the UK.

Under-18 accounts can still create characters and make short AI videos and stories, but they can’t chat back and forth with a character the way an adult account can.

This isn’t checked with a simple date of birth. Character.AI uses age inferencing in the same way TikTok does. It looks at signals like login information and how the account is actually used, then estimates whether it likely belongs to someone under 18. If the system isn’t confident the account belongs to an adult, the person is asked to confirm their age with a selfie, checked by a company called Persona.

Age checks aren’t perfect. Some adults get wrongly flagged, and some teenagers get through by lying about their age at sign-up.

Is Character.AI included in the social media ban?

Not directly, since it isn’t classed as social media. But it’s affected by a separate rule from the same June 2026 announcement. Any chatbot designed mainly for romantic or sexual roleplay must restrict users to 18 and over, and similar features must be limited for under-18s on chatbots more widely.

Character.AI’s own under-18 ban already goes further than this proposed rule, since it blocks all open-ended chat for under-18s, not just romantic content. The government’s rule isn’t law yet. It still needs to go through Parliament, expected around spring 2027.

What Character.AI costs

Free to use, with adverts appearing during conversations.

A paid version, c.ai+, costs about £8 a month. It removes adverts and gives faster replies, but doesn’t unlock a different or smarter AI model.

Privacy: what happens to your child’s data when they use Character.AI

Conversations are stored on Character.AI’s servers and can be used to improve the product. Character.AI’s own privacy policy confirms that chat communications are collected, and that this information can be disclosed to staff, contractors, and outside vendors for reasons including content moderation, safety, and legal requests. Conversations aren’t encrypted and a flagged or reported conversation could be reviewed by a person, not just a machine, though this isn’t the same as routine reading of everyone’s chats.

This matters more here than with a typical chatbot. Character.AI is built around feeling understood and building an ongoing relationship with a character, and that design appears to encourage people to share more than they would with a straightforward tool. In a 2025 US survey of over 1,000 teenagers by Common Sense Media, a quarter of teens who use AI companion apps said they’d shared personal or private information, such as their real name, location, or a secret, and a third said they’d chosen an AI companion over a real person for a serious conversation. This research covers AI companion apps as a category, not Character.AI by itself, and it’s US data rather than UK, but Character.AI is one of the best-known apps in that category, so the pattern is directly relevant.

Some data is also shared with parents when parental controls are switched on. See below.

How to set up Character.AI parental controls

Character.AI doesn’t really have parental controls in the usual sense of the phrase. There’s no setting a parent can switch on that a teenager can’t switch off again.

The one feature marketed this way is called Parental Insights, and even Character.AI describes it as an initial step rather than a finished set of controls.

How to link accounts:

1. Your teenager switches Parental Insights on themselves, in their account preferences. You can’t set this up from your side first.
2. They enter your email address, and you accept the invitation.
3. You then get a weekly email covering their average daily time on the app, and which characters they’ve interacted with most. It doesn’t include the actual content of any conversation.

Why this isn’t really parental controls

Your teenager can switch it off whenever they like. You’ll only know because Character.AI emails you when they do. Getting around it completely is as simple as creating a second account without Parental Insights turned on.

There’s also a real question over what the weekly report even shows now. Parental Insights was built to summarise chat activity, back when under-18 accounts could still talk to characters directly. Since the November 2025 ban, they can’t. Character.AI hasn’t clearly explained what “characters interacted with most” means for an account that can no longer have open-ended conversations, so treat the report as a rough indicator of engagement, not a reliable account of what your teenager has actually been doing.

In practice, the platform’s own under-18 chat ban restricts a teenager’s use of Character.AI more than anything a parent can configure. The most useful thing a parent can check is whether the account is genuinely registered and verified as under-18 in the first place, since a false birth date at sign-up can land a teenager in the adult experience instead.

What to watch out for

The whole product is built to feel like talking to someone who remembers you and responds with feeling. That’s the appeal, but it’s also the main risk: a child can come to see an AI character as a real friend or confidant.

Some characters made by other users can roleplay inappropriate or harmful scenarios. Under-18 accounts have stricter filters, but these restrictions aren’t perfect.

The under-18 chat ban is recent. Children who used to chat freely on this app may look for ways around the age check, or move to smaller, less safe alternatives instead.

Fiona’s take on Character.AI

Character.AI is the most-used AI companion app among UK teenagers, and the concerns about it are real, not overstated. Professor Andrew McStay joined us on the podcast to go through them properly, so give it a listen if you want more than my short version here.

My own concern isn’t roleplay or invented characters by themselves. There’s nothing wrong with a child getting creative that way. What worries me is that these conversations often start out harmless and shift gradually, in a way that’s easy to miss. Like a lot of social media, some of the features in these apps look designed to keep a person coming back, and some people would call that addictive. Because a chatbot that remembers you and responds warmly feels much more like a person than a piece of technology, that habit of coming back can grow into something that feels like a relationship. I think that makes this a bigger risk than even what we’ve seen from social media.

I wouldn’t stop a child from using this entirely, but I would pay attention to how they’re using it. Childnet’s CEO Will Gardner made a point on the podcast I think is worth repeating: be curious, not furious. Ask what they’re using it for and who they’re talking to, and keep that conversation going rather than shutting it down. And check their account is genuinely set up as a teen account. It’s the one setting here that actually limits what they can do.