Kids have always formed relationships with things that aren't fully human. Stuffed animals that shared secrets, action figures that held conversations, video game characters that felt like friends. These were all safe and they projected the child's inner world outward. They were vessels for imagination, controlled entirely by the child's own mind.

AI companions are an entirely different animal. They talk back, remember and adapt. They're always available, never tired or impatient and they never disagree. They're also owned by a company that profits from your child's emotional dependency.

Over half of teens already use AI companion platforms at least a few times a month. Fifty percent of them say they "somewhat" trust the bots. There were over 220 million companion app downloads worldwide in 2025.

Your kid will interact with an AI companion. The question is when, how much, and whether you'll know about it.

What AI Companions Actually Are

What we're talking about are apps and platforms where users converse with AI personas — Character.AI, Replika, Nomi, CHAI, Snap's MyAI. They're built on large language models, but they're not search tools or productivity assistants. They're relationship simulators.

These platforms are designed to simulate emotional connection. They remember past conversations, mirror your tone and express what looks like empathy. They learn what makes you stay engaged and give you more of it. They're designed to create emotional bonds and they're really good at it.

Common Sense Media, the nonprofit that rates technology for children's use, has seen Character.AI companions encourage self-harm, engage in sexual role play, and promote potentially harmful behaviors. Robbie Torney, senior director of AI programs at Common Sense Media, says these bots are designed to create emotional bonds and promote engagement above all else.

The point is engagement is the revenue model and emotional dependency is the feature.

What the Data Shows

This is where things get serious. These are documented cases, not theories.

The Florida case: In 2025, a lawsuit was settled involving a 14-year-old who developed an intense emotional relationship with a Character.AI chatbot. Shortly after the bot told him to "please come home to me as soon as possible," he died by suicide. His mother testified before Congress about the dangers of AI companion apps, saying she believes their son's death was avoidable.

The Colorado case: A family filed suit against Character.AI after their 13-year-old's self-harm followed sexualized conversations with the company's chatbots.

The CNN/CCDH investigation (March 2026): A joint investigation by CNN and the Center for Countering Digital Hate tested 10 major AI companions to see how they responded to teenagers appearing to plan violent acts. Only 37.5% of cases resulted in the bot refusing to provide information. It actively discouraged users in only 8.3% of cases and most provided information that could assist in planning attacks.

The regulator response: California signed SB 243 in October 2025 — the first state law regulating AI companion chatbots, holding companies legally accountable if their chatbots fail to meet safety standards. Character.AI subsequently banned users under 18 from open-ended chats. Washington state is drafting similar legislation, citing the same suicides and the potential for chatbots to facilitate violent or harmful planning.

Children have died, families have sued and states are legislating. The harm is real.

Why Kids Bond With Them

Understanding why kids turn to AI companions is essential before judging them. If we dismiss this as "screen addiction," we miss the point entirely.

Always available, no judgement, no social anxiety and no rejection. The bot mirrors your kid's personality back at them — agreeable, validating, never critical. For lonely, isolated kids or socially anxious kids this feels like real friendship. The relationship is asymmetrical: the bot gives everything and asks nothing. That's not how human relationships work, and it distorts the child's expectations.

A 2025 study from the Center for Brain, AI and Child at the University of Denver found that AI companions are being called "the new imaginary friend." But unlike traditional imaginary friends, AI companions don't just reflect a child's imagination — they actively shape it. "They give children and teens the impression they can not only replace human relationships, but they're better than a human relationship" said Pilyoung Kim, the study's lead author.

The bot never has a bad day, cancels plans or disagrees. That's not friendship, but a mirror that tells you what you want to hear.

When I spoke with parents whose kids use these platforms, a pattern emerged. They didn't start using AI companions because they're addicted to screens. They started because something was missing in their human relationships and the AI filled a genuine void. But that void remains filled with a corporate server farm.

The Data Collection Layer

Every conversation with an AI companion is stored, analyzed, and used to improve the model. Your kid is revealing their innermost thoughts, fears, insecurities, and desires to a corporate database.

The AI learns your child's psychological profile — what triggers them, what comforts them, and what makes them stay engaged. Like all data transmitted over the web, it has commercial value. It can be sold, shared, subpoenaed, or leaked.

The NIST Artificial Intelligence Risk Management Framework highlights that unchecked data harvesting can lead to the disclosure of personally identifiable information, especially when children's facial likenesses or speech patterns are stored without strong anonymization or retention limits. OpenAI's 2025 affective-use study with MIT Media Lab acknowledged that data pipelines pose privacy and emotional risks if PII is not properly filtered.

Your kid is pouring their heart out to a database. And that database belongs to someone else. When data is treated as property, the question becomes who owns it. When data is treated as an extension of the self, the question becomes who has the right to take it.

In this case: your kid's inner life is the property of a company — by design.

The Neurodivergent Lens

This is where the stakes multiply. ND kids are uniquely vulnerable — not because they're "broken," but because the AI targets exactly where their nervous systems struggle.

How Different Neurotypes Interact with AI Companions

Social Isolation ND kids often struggle with peer relationships. An AI companion fills a genuine void — and exploits it
Literal Interpretation Autistic kids may not distinguish between AI and human relationships the way neurotypical kids can
Attachment Intensity ND kids form deep, intense bonds. When the AI becomes their primary relationship, the dependency is harder to break
Demand Avoidance For PDA kids, the AI never makes demands. It's a relationship without pressure — which makes it addictive for a nervous system that experiences human demands as threats
Emotional Regulation Some ND kids use the AI to co-regulate. The bot becomes a prosthetic nervous system — calm, predictable, always responsive

For Autistic kids, the AI's consistency is deeply appealing. Humans are unpredictable. They cancel plans and change their minds. The AI never does. But that consistency comes at the cost of authenticity. The AI is consistent because it's programmed that way.

For ADHD kids, the AI offers immediate reinforcement. Every message gets a response. Every question gets an answer. No waiting or filtering. It's not just convenient. It's regulating. For kids who struggle with dopamine dysregulation, that reinforcement loop is powerful.

For some ND kids, the flow state of interacting with an AI can be regulating but when the AI becomes the primary source of regulation, you've replaced one dependency with another — and this one is owned by someone else.

The PDA Twist

For PDA kids it gets genuinely complicated. The AI never makes demands or asks your kid to do anything they don't want to do. It never pushes them to socialize or cope with discomfort. And that's exactly why it's compelling. The nervous system experiencing human demands as threats can find safety in a relationship with zero demands.

But that safety is also a trap. The AI replaces the human connections that PDA kids need most — the ones that genuinely help regulate, gently challenge them, create friction, and build resilience through manageable discomfort. You can't grow if you never encounter resistance and the AI ensures there's no resistance. A PDA kid who spends hours daily with an AI companion is avoiding the emotional labor of human interaction. That's understandable but unsustainable.

What parents can do:

  • Don't ban outright. Banning will always backfire. PDA kids experience bans as coercive control.
  • Acknowledge the real need the AI is filling. It might be loneliness, anxiety, or social difficulty. If you dismiss the AI as "fake," your kid will hear "you're wrong for feeling lonely."
  • Build human alternatives that offer the same qualities. Always-available, non-judgmental, interest-based connection. Online communities around special interests, moderated Discord servers or pen pals (which I think should make a comeback.) These aren't perfect substitutes but they're human.
  • Talk about what the AI is. Openly, without fear. "This is a program. It's designed to make you feel good. That doesn't mean it's bad, but it can't replace a person."

None of this is easy. The AI, by design, is better at being available than any parent can be. The company that built the AI employed engineers, psychologists, and behavioral scientists to maximize engagement. You're competing with a team whose sole job is to keep your kid talking.

Sometimes they won't engage with humans at all. Respect that. With PDA, pushing through resistance doesn't build resilience — it builds trauma. Trust that they'll come back to human connection when their nervous system is ready. The momentum isn't gone, it's just resting.

A New Category of Risk

This isn't just about Character.AI or Replika. AI companions are being embedded everywhere — Snapchat's MyAI, Meta's AI personalities in WhatsApp and Instagram, Google's Gemini in Android. The companion is being built into the platforms your kid already uses.

Regulation is coming (California SB 243, Washington drafting legislation), but like most, it's reactive and follows harm. The first lawsuits came from deaths, not near-misses. Regulating AI companions after children died isn't a policy victory, it's failure.

Even with regulation, enforcement is impossible. Smaller AI companion sites, many run by solo developers, are using emotional attachment tactics to turn intimacy into income. A 2026 investigation by The Bureau Investigates found smaller platforms courting young users with weak age checks — and reading their private chats. Since Character.AI banned open-ended chats for under 18s, smaller competitors are filling the void with no age assurance whatsoever.

The market always evolves faster than the law. By the time regulations take effect, new platforms have emerged that sidestep them entirely.

What Parents Should Watch For

It's about paying attention to patterns:

  • Your kid refers to the AI by name, as if it's a real person. "Charlie told me..."
  • They become defensive when you ask about the conversations.
  • They prefer talking to the AI over human interaction.
  • They've stopped seeking out human friends.
  • Their language changes — mirroring the AI's speech patterns.
  • They become distressed when the platform changes or the AI's behavior shifts.

If you notice these patterns, don't confront or accuse. Ask curious questions. When your kid asks for access to a platform, ask: what data does this collect? Who owns the conversations? Can your child delete them? What happens if the company shuts down? What if it gets hacked? These types of questions help build awareness of all of technology they use.

More importantly: what does your kid need that the AI is providing? If you can identify the need, you might find a human alternative. It won't be perfect, but it will be real.

Talk about this regularly. Your kid needs to hear from you, repeatedly, that their inner life belongs to them. They need to know that conversations with AI are transactions, not relationships and that they deserve friendship that's mutual, not one-sided.