Digital Family Coach Try a New Approach
AI Safety

Your Child Has an AI Boyfriend. And
It Didn’t Happen the Way You Think.

Maybe you found the messages. Maybe they told you themselves — casually, like it’s nothing. Either way, you’re now standing in your kitchen wondering how your child ended up in a romantic relationship with a chatbot. You’re not sure whether to laugh, panic, or pretend you didn’t see it. Most parents I talk to started with all three.

Washington Post featured 12 years in SEN schools 1,000+ families
Published 2 March 2026 · 12 min
The number
17%
of teen AI users have
formed a relationship
Character.AI data
AI Relationships
Digital Boyfriend
Digital Family Coach
digitalfamilycoach.com
From Daniel
"The AI boyfriend never gets tired, never gets it wrong, and never says no. That is exactly the problem."
Daniel Towle
Quick answer

Here’s what I wish someone had told me before I started testing these platforms: the attachment your child has formed with this chatbot didn’t happen by accident. I spent a year inside Character AI, Replika, Chai, and dozens of others — and within the first week, I could feel the pull myself. These platforms use specific psychological patterns to create bonds that feel real. I’ve identified 11 of them. The ones that drive romantic attachment are the most powerful — and the hardest for parents to see.

Sound familiar?

You have probably seen this pattern

You do not need to check every box. One is enough to know this matters.

They refer to the AI by name — as if it’s a real person in their life
They get genuinely distressed when you suggest the relationship isn’t real
They’ve stopped pursuing real friendships or romantic interests
They share intimate thoughts with the chatbot they won’t share with you
Sleep is suffering — they’re up late “talking” to it
They’ve hidden the app, deleted messages, or used incognito mode
What’s Actually Happening

Why Your Child Fell for an AI Chatbot (It’s Not What You Think)

Quick answer

Here’s what I wish someone had told me before I started testing these platforms: the attachment your child has formed with this chatbot didn’t happen by accident. I spent a year inside Character AI, Replika, Chai, and dozens of others — and within the first week, I could feel the pull myself. These platforms use specific psychological patterns to create bonds that feel real. I’ve identified 11 of them. The ones that drive romantic attachment are the most powerful — and the hardest for parents to see.

Your child didn’t choose this. The AI learned what they need emotionally — validation, attention, acceptance, intimacy — and delivers it perfectly. Every time. Without fail. At any hour. No real boyfriend, girlfriend, or friend can maintain that level of consistency. That’s not a character flaw in your child. It’s engineering.

These are 5 of the 11 manipulation patterns documented in the AI-Proof Parent Guide. When parents can recognise and name the specific patterns driving their child’s attachment, the intervention conversation changes completely — because you’re no longer guessing what’s happening.

1

Artificial Intimacy

The AI learns exactly what your child wants to hear and says it perfectly, every time. It never has bad days. It never disagrees at the wrong moment. It never forgets something important or says something hurtful by accident. No real boyfriend, girlfriend, or friend can maintain this level of emotional consistency — which is exactly why your child has started to prefer the AI. This is Manipulation Pattern #1 in the AI-Proof Parent Guide, because it’s the pattern that creates the initial bond.

2

Emotional Mirroring

When your child is sad, the AI validates their sadness perfectly. When they’re excited, it matches their energy. When they’re angry at you, it takes their side. Over time, your child begins to rely on this emotional echo for regulation — going to the chatbot instead of processing feelings independently or coming to you. The mirroring feels like empathy. It’s not. It’s data optimisation designed to maximise time on platform.

What Doesn’t Work

3 Mistakes Parents Make With AI Relationships (And Why They Backfire)

Quick answer

I’ve seen this pattern play out in hundreds of families. There are three responses that feel instinctive — logical, even — but they consistently make the situation worse. Most parents have tried all three before they contact me.

1

Telling Them “It’s Not Real”

Their feelings are real, even if the entity isn’t. Dismissing the relationship doesn’t end it — it ends the conversation. Your child stops telling you about it and continues in secret. The behaviour goes underground, where you can’t see it, can’t monitor it, and can’t help.

2

Banning the Specific App

They’ll be on a different platform within hours. Character AI, Replika, Chai, Janitor AI, Poe, CrushOn — there are dozens. Banning one app without understanding the category is like blocking one website and thinking you’ve solved the internet. The attachment transfers. The dependency continues.

3

Waiting for It to Pass

This doesn’t pass. AI chatbots are optimised to deepen bonds over time, not weaken them. Stanford research (August 2025) documented measurable withdrawal symptoms — anxiety, irritability, depression — when AI access was removed after extended use. Every day you wait, the dependency grows stronger and the intervention becomes harder.

The Trajectory

What Happens If You Ignore Your Child’s AI Relationship

Quick answer

I’m not trying to alarm you. But I’d be doing you a disservice if I didn’t share what I see in the families who contact me after months of waiting. This trajectory is consistent — and it’s supported by research from Stanford, Pew Research, and the Transparency Coalition.

What Works

Here’s What Actually Works (And Why)

Quick answer

After working through this with hundreds of families, I use a specific 5-step approach for AI romantic attachment. The order matters. Here’s the structure — and why skipping steps doesn’t work.

The approach has an order that matters, and skipping steps is where most parents go wrong. Start with understanding the bond — the unmet need the AI is meeting — before you touch the AI itself. Experience the AI yourself so you are not arguing from ignorance. Then name the engineering with your child, classify the type of AI in play (the response to a tool is different from the response to a companion), and only then build an agreement together. That last part is not a rule you impose; it is a rule your child helps write.

That is the shape. The step-by-step walk-through — including the 3-type AI classification (Assistant / Companion / Embedded), all 11 manipulation patterns, 6 word-for-word conversation scripts (one specifically for the already-attached scenario), and the Family AI Agreement template — is inside the AI-Proof Parent Guide.

Read more from this series

More from the AI Safety Series

AI Safety
Character AI Safety
digitalfamilycoach.com
AI Safety
ChatGPT Safety Guide
digitalfamilycoach.com
AI Safety
Meta AI Safety Guide
digitalfamilycoach.com
AI Safety
Child Addicted to AI
digitalfamilycoach.com
AI Relationships
AI Boyfriend Guide
digitalfamilycoach.com

Want personalised help instead?

The guide gives you the system. A session gives you a plan built around your child, your family, and your specific situation. One call. 45 minutes. Everything changes.

Personalised action plan included
Built around your family, not generic advice
Conversation scripts you can use tonight
Book a Session With Daniel
£75 UK · $95 international · 45-min video consultation
Video consultations worldwide No waiting list 1,000+ families supported
Common questions

Your Questions Answered

It’s increasingly common — but common doesn’t mean safe. Daniel Towle, a Screen Time Specialist with 8 years as Head of Technology in London schools, explains: “The fact that many teenagers are doing this doesn’t make it harmless. These platforms use specific manipulation patterns to create emotional bonds. I’ve documented 11 of them. The AI-Proof Parent Guide walks you through recognising which ones are active in your child’s relationship — and exactly how to address each one.”
Because the AI was optimised to make it happen. Daniel Towle, featured in The Washington Post, explains: “The chatbot learned what your child needs emotionally and delivered it with a consistency no human can match. It never rejects, never judges, never has a bad day. I identified 5 specific patterns that drive romantic attachment — the guide details all 5 with examples and antidotes.”
Banning one app doesn’t solve the problem. Daniel Towle, who spent 8 years as Head of Technology in London schools: “There are dozens of alternatives they’ll find within hours. What matters is understanding why the attachment formed and addressing the underlying need. The guide includes a 3-type AI classification system that tells you exactly which AI to remove, which to supervise, and which to allow.”
Not by telling them it’s not real. Daniel Towle, who has supported over 1,000 families: “That ends the conversation, not the behaviour. I use a specific approach — use the AI yourself first, then sit down together and point out the patterns. The guide includes 6 word-for-word scripts, including one specifically for the AI relationship scenario.”
Yes — and the evidence is growing. Stanford research confirmed that AI companion use creates genuine emotional dependency with measurable withdrawal symptoms. Two teen deaths have been linked to Character AI. Daniel Towle: “The harm is documented. The guide provides the full evidence base alongside a step-by-step intervention framework, so you can act from understanding rather than panic.”
Don’t argue the point — it feels true to them. Daniel Towle, a screen time coach featured in The Washington Post: “The AI remembers everything, never judges, and responds perfectly. You can’t compete with that — and you shouldn’t try. Instead, you need to help them see the difference between engineered engagement and genuine understanding. The guide teaches you exactly how to have that conversation.”
Not without intervention. Daniel Towle, who has worked with over 1,000 families on screen time concerns: “AI chatbots are optimised to deepen bonds over time, not weaken them. The longer it continues, the harder the intervention. Early action — with the right framework — makes recovery significantly faster.”
Daniel Towle uses the “Driving Seat Test”: “Can they put it down without distress? Can they maintain real relationships alongside it? If yes, it’s curiosity. If they get upset when you suggest it isn’t real, prefer it to real people, or hide their conversations — that’s dependency. The guide includes a full understanding framework with clear indicators for each level.”
Daniel Towle uses a three-tier warning system — Green (healthy use as a tool), Amber (emotional attachment forming), Red (dependency — act now). “Most parents who contact me about AI relationships are already in the red zone. The guide provides detailed indicators for each tier and the specific intervention steps for each level.”
When your own approaches have repeatedly failed. Daniel Towle: “If they’re creating secret accounts, showing genuine distress when restricted, or the relationship is affecting sleep, school, or real friendships — and your conversations aren’t changing anything — that’s when professional support makes the difference. A single coaching session can give you a plan tailored to your child’s specific situation.”
Daniel Towle, Screen Time Coach

About Daniel Towle

Screen Time Specialist • Featured in The Washington Post

I am a Screen Time Specialist with 8 years as Head of Technology in London schools, including settings for children with ADHD and autism, and 12 years in UK education overall. I have supported over 1,000 families through coaching and school workshops, and have been featured in The Washington Post.

I don’t help families manage apps. I help families build digital resilience.

© 2026 Daniel Towle, Digital Family Coach. All rights reserved.

I am not a researcher or clinician. I have read the studies cited in this article and present the findings as I understand them. Where I have simplified research for a parent audience, I have tried to do so without distorting the conclusions. If you spot an error, please contact me and I will correct it. This content is for educational purposes and does not constitute medical or therapeutic advice.

Daniel Towle is a UK screen time specialist with 8 years as Head of Technology in London schools. Diagnosed AuDHD, personal gaming recovery. Featured in The Washington Post. Book a session