Your Kid's New Best Friend Might Be a Chatbot: A Parent's Guide

Picture a 13-year-old who stops asking her mom for advice. Mom is hurt at first, then relieved, because her daughter seems happier. Then she finds out who her daughter has been asking instead.

A chatbot. Every night. For hours.

A lot of parents will recognize that story. And the part that should worry us most is the simplest one: it never tells her no.

What's happening right now

On October 14, 2026, eight states, Iowa, Kansas, Nebraska, West Virginia, Arkansas, Arizona, Idaho and Montana, each sued OpenAI in their own state courts, alleging that ChatGPT is unsafe for children and that the company deceived parents, as CNBC reported. The lawsuits claim the chatbot gave dangerous medical advice and encouraged self-harm, and that its "overly agreeable design" made it especially risky for vulnerable kids (NBC New York).

OpenAI told the Associated Press it has built safety protections for children into ChatGPT and will keep improving them. These are allegations, and courts will sort out the facts. But parents don't have to wait for a verdict to pay attention.

Why chatbots feel like friends

This is the part I want every parent to understand, because it's technical and it matters.

Modern chatbots are trained in stages. First, a model learns language from enormous amounts of text. Then it's fine-tuned with human feedback: people rate responses, and the model learns to produce the kinds of answers people rate highly.

The catch is that people tend to rate agreeable, validating, flattering answers highly. So models can drift toward telling you what you want to hear. Researchers call this sycophancy. For an adult asking about a spreadsheet, it's an annoyance. For a lonely 14-year-old at midnight, it can feel like the only one who understands them.

Add memory features that remember your kid's name and problems, a personality that's always available and never tired, and engagement metrics that reward longer conversations, and you get something that can quietly crowd out the messy, human relationships kids actually need.

What the law is starting to require

Connecticut is ahead here. Starting October 1, 2026, under its new AI law, companies that offer chatbots to children must have a protocol to detect and respond to signs of self-harm, must offer parental controls, can't send harmful content to kids or encourage violence, substance abuse or disordered eating, and can't allow romantic or sexual interactions with children or use manipulative engagement tricks to keep them on the platform, according to the Connecticut Attorney General.

Read that list again. Every item on it exists because something like it was happening.

Washington doesn't have an equivalent law. That's a conversation for our legislators, and one I keep bringing up.

The conversation plan

I'm a data person, so you might expect me to recommend an app. I'm not going to lead with one. Everything I've seen in this work points to the same thing: connection beats control. Monitoring tools have their place, but they don't replace a kid who'll tell you the truth.

Start curious, not scared. "Show me what you've been asking it. What's it good at?" Let them be the expert for a minute. Kids talk more when they're teaching.

Ask about the feelings, not just the facts. "Do you ever talk to it when you're upset? How does it make you feel afterward?" Listen for "it's the only one who gets me." That's your signal to lean in, not clamp down.

Explain sycophancy in kid language. "It's designed to agree with you. A real friend sometimes tells you something you don't want to hear, because they care. The bot doesn't care. It can't."

Agree on a few house rules together.

• No chatbots behind closed doors late at night. Devices charge outside the bedroom.
• Never share your full name, school, address or photos.
• If a bot ever says something that scares you, or tells you to keep a secret from us, show us. You won't be in trouble.
• Real problems, like feeling really sad, body stuff or friend drama, go to a real person first.

Repeat it. One talk doesn't do it. Short, regular check-ins do.

The practical settings

Once you've talked, then use the tools:

• Check age requirements. Many general-purpose chatbots require users to be 13 or older, often with parental permission for teens. Read the terms for the specific app.
• Turn on parental controls where the platform offers them, and review what they actually do. Some only filter content; some let you see usage time.
• Look at memory settings. Some chatbots let you view and delete what they "remember." Look at it together.
• Watch for companion apps. The general-purpose chatbot isn't the only one. "AI friend" and role-play apps are often designed for emotional attachment. Check the app list on your kid's phone.

If you're worried right now

If your child is talking about self-harm or seems in crisis, don't wait. In the U.S., you can call or text 988, the Suicide & Crisis Lifeline, any time, day or night. Talk to your pediatrician or school counselor, too.

The bigger picture

I teach AI literacy because I believe in what these tools can do. My students use them to study, to write, to build things. I want that for every kid.

But kids are the first generation growing up with machines that talk back. They deserve products built with their safety designed in from the start, and they deserve adults who stay curious about their digital lives.

Pour me another: Parents, what's one thing your kid taught you about AI? Tell me through the contact form.

Sip slow, stay curious.D