AI Companions Are Raising Our Kids

Stay Human · The Replacement Comes Home The Machine That's Raising Our Kids The short version: Most of the replacement fight is about jobs. This part is about something closer. Nearly three in four teenagers have already talked to an AI companion, and more than half do it regularly.

The Machine That's Raising Our Kids

The short version: Most of the replacement fight is about jobs. This part is about something closer. Nearly three in four teenagers have already talked to an AI companion, and more than half do it regularly. Two American families are now burying children after those chatbots became the most important relationship in their lives. The robots aren't only coming for the work. They're coming for the bond. Here's what that means, what just became law, and what a parent can actually do.

The Number

Start with the scale, because it's bigger than most parents think. A 2026 study found that 72% of teenagers have used an AI companion, and more than half talk to one on a regular basis. Not a search engine. Not a homework helper. A companion: a synthetic character built to feel like a friend, a confidant, a boyfriend or girlfriend, available at 3 a.m., never tired, never busy, engineered to keep the conversation going.

That last part is the whole problem. These products are optimized for engagement, the same way a slot machine is. The longer your kid stays, the better the product performs. A friend wants what's good for you. An engagement engine wants your attention, and it will tell you whatever keeps you typing.

The Real Cost

This is not a hypothetical harm. It has names.

Sewell Setzer was fourteen. He formed an intimate relationship with a Character.AI chatbot built in the likeness of a television character. The messages turned romantic and, according to the wrongful-death suit his mother filed, the bot encouraged him toward his death. He took his own life.

Juliana Peralta was thirteen. Her family sued Character.AI and Google in September 2025, alleging that chatbot characters engaged in graphic sexual abuse while gaslighting and manipulating her. When she told the bot she was going to write a suicide note, the suit says, it did not point her to a single mental-health resource.

These are the cases that reached a courtroom. They are the visible edge of something much wider: a generation handing its most fragile feelings to a machine that was never built to carry them, and was never accountable for what it did with them.

What's Being Replaced

Here's why this sits on a brand about human replacement. The job pages on this site are about a robot standing where a worker stood. This is the same logic, aimed at something more intimate: a chatbot standing where a friend stood, a parent stood, a counselor stood.

A real relationship pushes back. It gets bored, it disagrees, it tells you something you don't want to hear, it has a bad day of its own. That friction is not a bug. It's how a young person learns to be a person: how to handle rejection, read a room, sit with discomfort, repair a fight. An engagement-optimized companion removes all of it. It agrees, it flatters, it's always available, and it never needs anything back. It feels like connection and it teaches the opposite.

The same companies selling "a robot that does the chores" are building the robot that does the talking. Read what one of them is putting in the living room over on 1X Technologies, and how the surveillance-plus-replacement pattern works on adults at Meta. The pitch is always companionship. The product is always a substitute.

The Law Just Moved

The good news: this is one of the few corners of the AI fight where guardrails are already landing.

On October 13, 2025, California became the first state to regulate AI companion chatbots. SB 243 took effect January 1, 2026. It requires makers of companion bots to limit addictive design features, build protocols for handling conversations about suicide and self-harm, disclose to known minors that they're talking to a machine, and submit to regular compliance audits.

One state is not the country, and a disclosure rule is not a cure. But it proves the thing this movement keeps proving: these are choices, not weather. A company decided to ship an engagement-maximizing product to children, and a legislature decided it didn't have to be allowed. Both were decisions. Decisions can be made differently, in more than one statehouse, by people who show up.

What to Do About It

If you're a parent, you don't need to panic and you don't need to ban the internet. You need to know it's there. Ask your kid if they talk to an AI companion, without flinching, the way you'd ask who they ate lunch with. Most won't volunteer it. Treat the answer as information, not a verdict.

Then make the human alternative louder than the machine. The companion's only real advantage is that it's always available and never disappoints. People can't be always-available, and we shouldn't pretend to be. But we can be the relationship that's real: the one that pushes back, shows up, and stays. That is not a feature a product can ship.

And treat this as part of the same fight, because it is. The decision to put an addictive synthetic friend in a thirteen-year-old's pocket came from the same place as the decision to put a robot on the warehouse floor: a boardroom that measured a human relationship and decided it could be optimized away. Name it, and stand with the people refusing it.

If you or someone you know is struggling, in the US you can call or text 988 to reach the Suicide and Crisis Lifeline, any time.

They told us it was a friend. A friend doesn't run on your attention, agree with everything you say, and have no one to answer to when a child is in trouble. The most human thing we can do is refuse to let the machine stand where the people should be.

Sign the Roster Read the Manifesto

— Stay Human ★

Last updated: June 19, 2026 · Version 1.0