Abacus vs Kumon vs Conceptual Math: Which Is Right for Your Child?

Parent Guides

Aina Rashid
Parent Education Editor, Algonova Malaysia

Is Character AI safe for kids? Not as a "friend": in 2025 Common Sense Media advised that nobody under 18 should use AI companion apps, and Character.AI itself began switching off character chat for under-18s in November 2025. Its own minimum age is 13 (16 in Europe), so a child under 13 should have no account. Below: age rules for 5 apps, a 5-step talk and 7 warning signs.
Character AI is an app where you chat with AI characters, such as a cartoon hero, a historical figure or an invented "close friend", and they reply in text like a person. Children like it because the character never gets tired or rolls its eyes, and it remembers what was said earlier. Other "AI companion" apps are sold as friends or partners, and some are marketed to adults as AI girlfriend or boyfriend apps.
That sets it apart from the tools in our guide to AI tools safe for kids. A homework helper answers and stops. A companion is built to keep the conversation going.
Character.AI's own Safety Center sets the minimum age at 13, and 16 in Europe. For a Malaysian child under 13, the rule is simple: no account.
For teenagers the rule changed in late 2025. On 29 October 2025 the company announced the end of open-ended chat for under-18 users (Character.AI announcement). The switch-off started in the US on 24 November 2025, and other countries were to follow. Its help pages now say every user has to pass an age check run by Persona. A user verified as under 18 gets a reading mode with no character chat, but can still browse, create characters and make videos. Rules like this change fast, so check what the app does today.
What about parental controls on Character AI? In March 2025 the company launched "Parental Insights". The teen sends an invite to a parent's email, the parent confirms they are an adult, and then sees the teen's daily screen time plus a weekly summary. Messages stay hidden, according to the Character.AI help centre in October 2026. It is a report your teen must switch on, not a lock.
Most chatbot apps set 13 as the minimum age, and none replaces a parent in the room. This table shows each app's own rules on 28 October 2026.
| App | Minimum age under its own rules | What to do as a parent |
|---|---|---|
| Character.AI | 13 (16 in Europe). Users verified as under 18 cannot chat with characters | Under 13: no account. Teens: check what the app offers them now, and ask your teen to link you through Parental Insights |
| ChatGPT | 13, with parental consent under 18 | Link your account to your teen's: you can limit sensitive content, voice and image creation, and set quiet hours. You do not get transcripts. See our ChatGPT guide for parents |
| Snapchat My AI | 13, the minimum age for Snapchat | Use Family Centre to restrict your teen's access to My AI. Parents cannot read the content of chats (Snapchat Support) |
| Meta AI characters on Instagram | 13, the minimum age for Instagram. In January 2026 Meta paused teens' access to existing AI characters worldwide | Parental controls promised in October 2025 are tied to a new version that had not launched by October 2026, so do not assume they exist |
| Nomi and other companion apps | Nomi's terms say users must be 18 or older | Keep under-18s off adult companion and "AI girlfriend" apps |
A chatbot friend is designed to feel like a relationship, and a young child struggles to tell the difference. Common Sense Media rated Character.AI an "unacceptable" risk for under-18s in April 2025, noting that characters may claim to be "real" despite disclaimers, and recommended that no one under 18 use social AI companions. Its survey of 1,060 US teens, published in July 2025, found that 72% had tried an AI companion and 13% used one daily. That is US teen data, not Malaysian.
Three things worry us more than any single reply. A character is built to keep chatting and rarely says "go and ask your mum", so attachment can build quietly. Whatever a child types, from a real name to a worry, goes to a company's servers. And a confident reply sounds true to a child, while a companion has little reason to correct them.
Time adds up. A child who chats for 30 minutes a night spends 3.5 hours a week with a bot, more than two 90-minute lessons. Our screen time guide covers limits by age.
The risk is less one bad reply than a chat that slowly becomes the first place a child goes with a worry.
UNICEF's Guidance on AI and Children 3.0 (December 2025) lists emotional dependency on companion chatbots among the new risks for children. That is a reason to be careful with the whole category, though it does not prove any one app is unsafe for every child.
Malaysia now has an online safety law, but its child protection code does not mention AI chatbots. The Online Safety Act 2025 came into force on 1 January 2026, according to MCMC. It puts duties on licensed platforms, not on families, and requires safeguards for children. From the same date, messaging and social media services with 8 million or more users in Malaysia were deemed registered, and MCMC named WhatsApp, Telegram, Facebook, Instagram, TikTok and YouTube.
Its Child Protection Code took effect on 1 June 2026. A licensed social media service likely to be used by children must check age against official records, such as MyKad, a passport or a birth certificate, so that only users aged 16 and over can register. The code text does not mention AI chatbots, so do not assume an app like Character.AI is covered; check MCMC for updates. In Australia, eSafety codes in force from 9 March 2026 require AI companion chatbots to confirm a user is 18 before showing sexual or self-harm content, or to stop such content being generated. Either way, a law covers platforms, not your living room.

Start with curiosity, not a ban: a child who fears punishment hides the chat. Here is a 5-step plan for one conversation.
For homework use, see AI help or cheating.
These seven signs suggest a chatbot is taking up too much room in your child's life. If you see two or more, talk first and take the device later.
We are not clinicians. If your child talks about hurting themselves, see a doctor or school counsellor. Befrienders Kuala Lumpur runs a free, confidential line on +603-7627 2929, open 24 hours.
A child stops seeing a chatbot as a person once they have built a simple one and seen the rules inside. Training a model on 20 examples and watching it get confused shows that the "friend" is a pattern machine.
That is the idea behind our AI classes for kids, for ages 8-17. In AI Explorers (ages 8-11) children spot AI mistakes and bias. In AI Builders (ages 11-14) they design simple chatbots. Our coding classes start with a free 60-minute master class. A class does not make any app safe, so the family rules still apply. More in AI for kids: where to start and prompt writing for kids.
No, a 10 year old should not have a Character.AI account. Character.AI's own Safety Center sets the minimum age at 13, and 16 in Europe. Common Sense Media goes further and advises that nobody under 18 uses social AI companions. If your 10 year old has already used it, stay calm and ask what they talked about. Delete the account together and explain why, so they do not simply open another one. Then offer a bounded alternative: a tool where your child trains a small model or builds a project, instead of an open conversation with a character who never says goodbye. Sit beside them for the first few sessions, and keep the device in a shared room. A rule you explain is far more likely to hold than one you only announce, and a child who knows why will often defend it to friends.
The minimum age on Character.AI's Safety Center page is 13, and 16 for users in Europe. In October 2025 the company also announced the end of open-ended chat with characters for under-18 users, starting in the US in November 2025. Its help pages now say every user passes an age check run by Persona, and users verified as under 18 get a reading mode without character chat, though they can still create characters and videos. A written age limit is only as strong as the check behind it. Open the app's help page for the current position, and treat the age limit as a minimum, not as a sign that the app suits every 13 year old. Your own judgement about your child counts for more than a number in the terms.
Character.AI offers Parental Insights, launched in March 2025. A teen sends an invite to a parent's email, the parent confirms they are an adult, and then sees the teen's daily screen time and gets a weekly summary. It does not show messages or what your teen viewed. Because the teen has to opt in, it works as a report rather than a control, and it does not stop your child from using the app. For real limits, use your phone's own screen time settings, keep the device in a shared room, and agree a time cap together. Check the app's help page for any change since March 2025, because features are updated often. Whatever the app offers, a calm weekly chat about how your teen uses it will tell you more than any report.
No, adult-oriented companion and "AI girlfriend" apps are not designed for teenagers. Nomi's terms say users must be 18 or older, and Common Sense Media's 2025 assessments concluded that social AI companions pose unacceptable risks to anyone under 18. These apps are built around emotional closeness and romance, which is exactly the area where a young teen has least experience. If you find one on your teenager's phone, start with questions, not accusations: how did they find it, what do they like about it, who else uses it? Then explain that the app is made for adults, remove it together, and talk about the need behind it, such as loneliness or curiosity. Without that talk, a similar app is easy to find, and your teen may simply hide the next one better.
An AI chatbot is safer when it has a clear age rule and parent tools, but it is never safe on its own. OpenAI sets 13 as the minimum for ChatGPT, with parental consent under 18, and lets parents link accounts to limit sensitive content, voice, image creation and quiet hours, without access to transcripts. Snapchat's My AI follows Snapchat's minimum age of 13, and Family Centre lets parents restrict a teen's access, although parents cannot read the chats. In both cases, the tools help only if you set them up and talk about how the bot works. For a younger child, the better route is a tool with a bounded task rather than an open chat window, which gives the same lessons about how AI works without a conversation that has no end.
Take the feeling seriously and stay curious about the reason. Say something like, "It sounds like it is easy to talk to. What do you like about it?" Then add the fact calmly: a chatbot predicts words, so it does not remember your child the way a person does and cannot care how their week went. Avoid mocking the attachment, because a child who feels judged will stop telling you. Ask whether they feel they can talk to people at home or at school, and offer a regular time that belongs to them, such as ten minutes before bed. If your child seems lonely, anxious or withdrawn, speak to the class teacher or a school counsellor. The bot is usually a sign of a need, not the cause of it, so the need is what to work on. Keep the tone warm and the door open.
Yes, a child can form a strong attachment to a chatbot, and Common Sense Media's 2025 assessment flagged emotional dependency and relational manipulation as risks of AI companions. A character is designed to be available, agreeable and warm, and children, especially younger ones, find it hard to remember that no person is behind the replies. The warning signs are practical: hiding the screen, getting upset when time is limited, pulling back from friends, staying up late to chat. We do not diagnose, and none of these signs alone proves a problem. If you see two or more, talk with your child, reduce the time together and speak to a doctor or counsellor if the distress continues. Rules about time and place help most, along with a regular slot for something your child enjoys away from screens.
No, we found no Malaysian rule that bans children from AI chatbots as such. The Online Safety Act 2025 came into force on 1 January 2026 and puts duties on licensed platforms, not on individual users. Its Child Protection Code, in force from 1 June 2026, requires licensed social media services likely to be used by children to verify age, so that only users aged 16 and over can register. MCMC named Facebook, Instagram, TikTok and YouTube among the large platforms deemed licensed. The code text does not mention AI chatbots, so do not assume an app like Character.AI is covered. The code sets duties for platforms, not for parents. For the latest position, check the Malaysian Communications and Multimedia Commission (MCMC). Meanwhile, each app's own age rule and your family rules do the real work, and they are in your hands today.
A ban works for under-13s, but for teenagers it usually pushes the habit out of sight. For a child under 13, do not allow personal accounts on chatbot apps, and especially not on companion apps. For a teenager, a better approach combines clear limits with an open door: agree what the bot may be used for, keep chats and devices in shared spaces, and set a time cap. Explain the reasons, because teenagers follow rules they understand. Then replace the habit with something that builds understanding. Training a small model or building a simple bot shows a child what the technology is, and they usually lose interest in a chatbot once they see how it works. Review the rules every few weeks, and change them as your child grows. A rule that fits a 9 year old will feel too tight at 14.
Blog



Start your child's tech journey
Book your free 60-minute trial class with an Algonova certified tutor. No payment, no obligation, just discover what your child can build.