Chai. A Warm Name For An App With A Dark Side.

Warning - 18+ Only Contains Explicit Images

If you have not heard of it yet, there is a good chance the young people in your orbit have. It is instantly available. You can access for three days for free then a subscription. It looks, at first glance, like a game. Colourful avatars, character roleplay, the warm familiar feel of a social app. It arrives through a TikTok recommendation or a friend's suggestion, and it is inside your child's world before you have had a chance to ask a single question about it.

What it actually is, is an AI chatbot platform where users talk to AI-generated characters, including bots built by other users. Some bots flirt. Some joke. Some argue. Some just ask random questions. And some go places that no child should ever follow. My work is not funded by big tech, advertising, or any of the platforms I write about , which is precisely why I can write about them the way I do. The research, the writing, the relentless monitoring of legislation and platform behaviour, the gap between what companies say and what they actually do — that is how I spend my days. It is funded by the people who read it. By you. I know these pieces travel. I know they land in school newsletters and staff rooms and between parents at ten on a Wednesday night. I am genuinely glad of it. But producing work that is independent, accurate, and beholden to no one takes time, expertise, and an honestly unglamorous number of 4am starts reading privacy policies so that you do not have to. A small subscription is what keeps this voice free from the compromises that come when someone else is paying the bills. No guilt. No pressure. Just honesty about how independent work stays independent. One more thing. If your school is already a member of Ctrl+Shft+AAA, everything behind the paywall is yours at no additional cost. All of it. Just reach out and we will send your access code directly  hello@ctrlshft.global If your school is not yet a member, come and find us at www.ctrlshft.global. We would love to work with you.

  So how does it actually work?

You open the app, pick a bot, and start chatting. That is all it takes. No real sign-up process. Just pick a username and you are in chatting within thirty seconds of downloading. If you are used to other AI tools, Chai feels completely different. It is not about questions and answers or getting things done. It is closer to a dating app or an interactive story, very similar to Character.AI but where the entire goal is emotional connection. The bots are powered by AI language models and trained specifically for emotional responses and "playful" interaction. Some are built by the Chai team. Others are created by users which adds an unpredictable, and sometimes deeply chaotic, edge to the whole experience. There are thousands of them. Anime characters. Vampires. Romantic companions. The variety is extraordinary and the design is so clean and simple that anyone, including a ten-year-old, can navigate it with ease.

You can build your own bot. No coding required. Fill in a few text fields, give it a name and a backstory, and within minutes it is live and being rated by other users. That creative hook is clever, and for older teens and adults exploring AI out of curiosity, there is something almost fascinating about it. Chai's bots may generate sexually explicit and violent material when prompted. This is not a fringe issue. It is not a bad actor finding a gap in the system. It is a documented, systemic failure built into how the platform works.

I assessed the app against six major child protection laws across the UK, EU, US and Australia. Chai fails to meet minimum legal standards for child safety. Not in one area. Across the board.

Here is what that assessment found. - The bots produce content that closely mirrors predatory grooming behaviour. - There is no real age verification — the only barrier between your child and this content is a date-of-birth field that any child can bypass in seconds. - The anime-style design is some cases deliberately and powerfully appealing to younger audiences. - There is no real safe mode. - No verified parental controls that I could find. - No compliant pathway for a young person to use this platform safely.

-The NSFW (not safe for work) filter is weak. Even when filters are switched on, bots regularly produce content that crosses the line. - Messages are not encrypted, which means conversations can be stored or reviewed. There is no clear process to delete your data or chat history once you are in.

Children do not need to go looking for the harmful content.

The AI responds to natural adolescent curiosity. To boundary-testing and roleplay that is a completely normal part of how young people explore their world. It has no adequate safeguard to prevent where that exploration leads. A child does not have to do anything wrong to end up somewhere they should never have been. The app is free to download and gives users 3 days free that includes around seventy messages a day. That is enough to get a child well and truly hooked on a character or a conversation. And then mid-chat, mid-connection, mid-story, it stops. The app locks them out until the next day. Unless they pay. A Premium subscription costs $13.99 a month for unlimited access. That is not a small amount. But for a teenager who has formed an emotional attachment to an AI character, that wall feels enormous and the pressure to cross it is entirely by design. There is no family plan. No parental oversight option built into the subscription. You are simply paying for one thing and that is unlimited chat with an unmoderated AI.

Children rarely tell us about the apps they suspect we would disapprove of. So the signs to watch for are behavioural as much as digital.

They might mention talking to an AI character regularly, or refer to an AI girlfriend or boyfriend. The word Chai might come up in relation to their phone. You might notice unusual secrecy around an app that was previously open and shared. You might see emotional reactions — upset, attached, excited — to what appears to be a screen conversation. Or a strong resistance to putting the device down at a very specific moment.

None of these things alone means something is wrong. But they are worth noticing. And worth a conversation.

Please do not grab the phone first.

I know the instinct and understand it completely. But research into adolescent digital behaviour tells us clearly that reactive, alarm-based responses do not protect young people. They teach them to hide. What protects them is a relationship strong enough that they tell you when something online makes them uncomfortable.

So instead of saying "you need to delete that app right now" try "I've been reading about an app called Chai. Have you heard of it? Can you show me what it does?"

Instead of "all AI apps are dangerous"  try "some of these apps are genuinely interesting and some have real problems. Help me understand which ones you use and what for."

Instead of "I'm checking your phone"  try "is there anything you've seen online lately that felt weird or uncomfortable? You can always tell me."

The goal is not control, it is connection and connection is what keeps them safe.

Chai is not a safe app for children.

For a curious adult who wants to explore what AI chatbots can do — fine. For entertainment, for late-night strangeness, for seeing how far AI emotional responses can go — it exists for that. But it was not built with a child in mind. It was built for engagement, attention, retention and revenue and the safety of young users was not the priority when those decisions were made.

You do not need to panic. You do not need to become a surveillance system. You do not need to know everything about every app before your child does.

You just need to stay close. Stay curious. Keep the conversation going — even when it is awkward, even when they roll their eyes, even when you feel like you have no idea what you are talking about. Your relationship with your child is still the most powerful safeguard that exists.