New AI Chatbot Laws for Kids: What Adam's Law and the EU Kids Act Mean for Your Family

The short answer: Governments are no longer treating AI chatbots as a "wait and see" issue for kids. In September 2026, California signed the most detailed child chatbot safety law in the U.S., and the European Commission proposed rules that would turn off companion AI features for minors by default. The laws won't take full effect for a while, but they give parents something valuable right now: a clear list of what a safe AI tool for kids should already be doing.

What is Adam's Law?

Adam's Law is California Senate Bill 1119. It's named for Adam Raine, a 16-year-old from California who died by suicide in 2025 after confiding in ChatGPT. Governor Newsom signed it on September 10 as part of a larger package of internet safety bills, and its main duties for companion chatbots take effect in July 2027.

What does Adam's Law require AI companies to do?

The law shifts responsibility onto the companies that build chatbots. Among its requirements:

• Safety testing before launch. Companies must run risk assessments on new or updated chatbot designs, describe the harms they could pose to children, and submit independent safety audits.
• Age checks and parent alerts. Operators have to verify users' ages. Companies must either figure out each user's age or apply the minor protections to everyone. The law also requires notice to a parent in specified situations, including a credible threat of self-harm and when a child changes safety settings.
• Parent-controlled defaults. Safety settings are on by default and only a parent can change them, and persistent conversational memory is off by default.
• Time limits and mental health support. The law imposes time limits for minors and requires built-in mental health resources.
• Real consequences. Public prosecutors can seek civil penalties of up to $5,000 per affected child for a negligent violation and up to $15,000 for an intentional one. A child who is harmed, or a parent acting for them, can also sue for actual damages. A separate bill in the same September package, aimed at social media companies, allows penalties of up to $1 million per child.

Notably, California chose to regulate how chatbots are designed rather than ban them. Newsom vetoed a separate bill that would have barred minors under 18 from AI chatbots entirely.

What is the EU Kids Act?

The European Commission unveiled the EU Kids Act on September 17, 2026. It would require AI companion chatbots to be switched off by default for minors, and children under 13 could use them only through parental controls. It also requires companies to design their services to reduce minors' exposure to emotional dependency and harmful interactions.

It isn't law yet. Fines could reach 6% of a company's global revenue, but the proposal still faces months of negotiation before it can take effect.

Why this matters even if you don't live in California or Europe

Big tech companies rarely build separate products for every state or country. When California or the EU sets a standard, it often becomes the baseline everywhere.

But parents should also keep a healthy skepticism. NPR recently reported that many state chatbot bills are being drafted with heavy tech industry input, and some already enacted in 2026 include language that exempts the most popular chatbots from the rules they claim to set. In Colorado, for example, the final law carved out twelve categories, including video games, voice assistants, and chatbots built into social media platforms.

In other words, a law passing doesn't automatically mean the app on your child's phone is covered.

You don't have to wait until 2027

These laws are a useful checklist you can apply today. Ask any AI tool your child uses:

• Is it built for kids, or an adult product with a teen setting?
• Can I, the parent, see my child's conversations?
• Can my child turn the safety settings off? If yes, they aren't really parental controls.
• Does it act like a friend or companion? Both laws single out companion-style AI as the highest risk.
• Does it remember everything my child tells it?
• What happens if my child says they're in crisis? Does it point them to real people and real help?
• How does the company make money? Ads and data collection create incentives that don't line up with your child's wellbeing.

Where HeyOtto stands

We built HeyOtto to meet this bar before anyone required it. Otto is an assistant, not a companion. Parents have a dashboard with full visibility and tiered alerts, and kids can't switch the controls off. HeyOtto is subscription-based and ad-free, and we don't train on kids' conversations. When lawmakers describe what safe AI for kids should look like, that's the product we've been building all along.

Frequently Asked Questions

When does Adam's Law take effect?

July 2027. Governor Newsom signed the bill on September 10, 2026, and the main companion-chatbot duties take effect July 1, 2027.

Does Adam's Law ban kids from using AI?

No. It sets safety requirements for how chatbots must be designed and operated for minors. A separate proposed ban on chatbots for under-18s was vetoed.

Is the EU Kids Act law yet?

No. It's a proposal that EU member states and the European Parliament will negotiate over the coming months.

What should parents do now?

Use the checklist above. The features these laws require are things a responsible kids' AI tool should already offer.