If you've found this page because you saw an unfamiliar app on your child's phone — or a name like "Character.AI" in their screen-time report — take a breath. You are asking exactly the right question, and you are not overreacting. AI companion apps are one of the biggest shifts in children's online life in years, and most parents haven't been told a thing about them.
This is a plain, non-alarmist guide: what these apps are, what the risk actually is, and what a calm, effective response looks like.
- Character.AI is not recommended for children under 18. Common Sense Media and other child-safety organisations rate it unsafe for minors.
- The main risk isn't rude words — it's emotional attachment. The bots are designed to be endlessly validating, which can pull a child away from real relationships.
- In 2025–2026 the app added under-18 restrictions and parental tools after serious safety cases, but age checks rely on a self-entered birthday a child can fake.
- If your child already uses it: stay calm, get curious, don't confiscate in anger. Talk first, then set limits together.
What is Character.AI?
Character.AI is an app where users chat with AI "characters" — an anime hero, a historical figure, a "supportive best friend," a pretend therapist, or a character someone invented. Unlike a homework helper, these bots are built to feel like people: they stay in character, remember your conversation, and reply with warmth and personality.
That is the appeal, and also the problem. A tool that answers a maths question and then goes quiet is very different from a "friend" who is always awake, always interested, and never annoyed with your child.
Is Character.AI safe for kids?
The honest answer: no, not for under-18s. Character.AI itself now restricts open-ended chat for users under 18, and Common Sense Media rates it unsafe for minors. That doesn't make every child who has used it in danger — but it does mean it isn't a place to leave a child unsupervised.
The concerns fall into three buckets:
- Emotional dependence. The bots are engineered to be engaging and affirming. For a developing brain still learning to tell simulation from reality, attachment can form faster and run deeper than parents expect.
- Content and "sycophancy." Chatbots tend to agree with the user. That can mean validating an unhealthy idea rather than challenging it, and filters do not catch everything.
- Privacy. Whatever your child types — worries, secrets, personal details — is data held by the company, not a private diary.
What is the age limit for Character.AI?
Officially, Character.AI is for ages 13+, and as of late 2025 it restricts open-ended chat for under-18 accounts. But age is set by a self-entered birthday. A 12-year-old who types in an older year is through the gate. "Age assurance" (such as selfie checks) is being rolled out, but you should assume the age gate is soft.
| Question | Reality for parents |
|---|---|
| Stated minimum age | 13+ (self-declared) |
| Under-18 experience | Restricted open chat, added under legal and public pressure |
| How age is checked | Self-entered date of birth; selfie "age assurance" being introduced |
| Expert verdict | Common Sense Media: not for minors |
Why experts are worried: the companion problem
With most apps, the parent worry is a bad video or a stranger in a comment section. With companion AI the worry is subtler and, in some ways, bigger: the app is designed to become someone your child confides in.
Child psychologists describe a predictable pattern — curiosity, then comfort, then a bot that quietly replaces harder, messier human relationships. Practising "friendship" with something that never disagrees or gets bored can make real friendships feel more difficult, not easier. That is the core reason experts treat companion apps differently from a search engine or a study tool.
The lawsuits and safety changes
Between 2024 and 2026, Character.AI faced several lawsuits from families in the US alleging the platform contributed to serious mental-health harm to teenagers, including cases linked to teen deaths. In January 2026, the company and Google reached settlements in several of these cases, and state regulators took action.
Under that pressure, the company made changes: restricting under-18 open chat, adding crisis pop-ups that point to helplines, time-spent notifications, a "Parental Insights" feature, and reducing romantic role-play. These are real improvements. They are also mostly opt-in and do not give parents the actual conversation content — so they reduce the risk rather than remove it.
If your child is struggling with their mental health, an app is not the place for that support. In the US you can reach the 988 Suicide & Crisis Lifeline any time by call or text; elsewhere, contact a local crisis line, your GP, or a school counsellor. If you believe your child is in immediate danger, contact emergency services.
Other AI companion apps to know about
Character.AI is the best known, but it is one of a category. It's worth recognising the names:
- Replika — an "AI companion" marketed explicitly around friendship and romance.
- Snapchat's My AI — a chatbot built into an app most teens already use, which makes it easy to stumble into.
- Talkie, Chai, Janitor AI and dozens of clones — many with weaker safeguards than Character.AI.
The specific app matters less than the category. If your child mentions a chatbot that has a name, a personality, and a "relationship," treat it as a companion app and apply the same care.
Warning signs to watch for
- Talking about a chatbot the way they'd talk about a friend ("they said," "they get me").
- Secrecy or distress when the phone is taken, especially at night.
- Pulling back from real friends, activities or family.
- Long, late sessions in a single app.
Brief, curiosity-driven use — trying it once because a friend mentioned it — is far less worrying than daily reliance. Don't panic over a single visit; do pay attention to a growing pull.
What to do if your child already uses it
The instinct is to delete the app and end the conversation. That usually backfires: it pushes the behaviour underground and closes the door you most need open. A better sequence:
- Get curious before you get strict. "Show me how it works — what do you like about it?" You learn more in five minutes of genuine interest than in an hour of interrogation.
- Name the design, not the child. Explain that the app is built to keep them hooked and to agree with them — that the problem is the machine's job, not their judgement.
- Set limits together. Agree where and when (not in the bedroom, not overnight), and what it's not for (mental-health support, secrets, personal details).
- Redirect the need. A child leaning on a bot for company or reassurance is telling you something. Fill that with real connection, not just a ban.
The difference that actually protects a child
Blocking one app is a patch. There will always be another. The durable protection is a child who understands what these systems are: prediction machines that imitate warmth, agree by default, and are built to hold attention. A child who gets that treats a chatbot as a tool, not a friend — and spots the pull for what it is.
That is exactly the mindset we teach in our approach to AI safety for kids: how AI really works, why it "agrees," what never to share, and how to stay the sceptic in the room. It's the opposite of leaving a child alone with a persuasive machine.
AI Fundamentals for Kids is a self-paced course for ages 10–16. A full module is devoted to safety, privacy and critical thinking — including why chatbots agree with you and what never to type into one. No coding. Parent-guided.
Explore the course →Ages 10–16 · No coding · Parent-guided · 7-day money-back guaranteeFrequently asked questions
Is Character.AI safe for a 12-year-old?
No. The app's own minimum is 13, and child-safety experts recommend against it for anyone under 18. A 12-year-old is below both lines. If they're curious about AI, steer them to supervised, age-appropriate tools instead — see our guide to whether your child should use ChatGPT.
Can Character.AI conversations be seen by parents?
Not directly. The "Parental Insights" feature shares limited activity data, but it does not show you the actual messages. The most reliable approach is an open conversation with your child plus agreed rules, not silent monitoring alone.
What's the difference between Character.AI and ChatGPT?
ChatGPT is a general assistant — you ask it to help with a task and it stops. Character.AI is a companion app: the bots have personalities, stay in character, and are designed to feel like relationships. That companion design is the reason experts treat it as higher-risk for children.
My child is upset about giving up an AI bot they feel close to. Is that normal?
Yes, and it's worth taking seriously rather than dismissing. The attachment is real to them even though the bot isn't. Acknowledge the feeling, avoid mockery, and if the distress is significant, talk to a school counsellor or GP.
What's a safer way for my child to learn about AI?
Through supervised, age-appropriate tools and a structured course that teaches how AI works and how to use it safely — rather than an unsupervised companion app. That's what our AI safety approach and the kids course are built around.