Google is letting kids under 13 talk to Gemini

Google is about to hand its flagship AI over to children under 13, and it is putting the burden of the talk on parents.

According to a report from The New York Times, Google has started emailing parents who manage Android devices through its Family Link parental controls. The message is simple: kids under 13 with managed family accounts will soon be able to access Gemini apps directly on their devices.

Kids can turn it on themselves

Google frames Gemini as a helpful companion for young learners. The company says kids can use the chatbot for homework help or to generate bedtime stories.

On the privacy front, Google is adopting the same stance it uses for its Google Workspace for Education accounts. Interactions from these under-13 accounts will not be used to train its AI models.

Interestingly, kids will be able to enable and access Gemini on their own without needing parental pre-approval on the device. Google spokesperson Karl Ryan confirmed that parents will receive an additional email notification the first time their child opens Gemini, and parents can disable access entirely through Family Link at any time.

Have the talk, Google tells parents

Putting an LLM into the hands of an elementary schooler comes with obvious safety concerns, and Google is not pretending otherwise.

In its email, Google warned parents that Gemini can make mistakes and that kids may encounter content their parents would rather they not see. Beyond harmless hallucinations like counting letters in words or suggesting glue as a pizza topping, conversational bots can blur reality for kids.

Services like Character.ai recently faced lawsuits and heavy scrutiny after young users struggled to distinguish chatbots from real people. In response, that platform had to roll out stricter parental controls and behavioral safeguards.

Google's defense strategy relies heavily on household conversations. The company explicitly advises parents to sit down with their kids, explain that Gemini is not a human being, and warn them never to share sensitive personal information with the chatbot.

Why it matters

Google is making a calculated push to capture the next generation of users before competitors do. School-age kids are already turning to AI for schoolwork, and Google would rather bring them into a monitored Family Link ecosystem than watch them sneak around on unmonitored adult accounts.

The promise not to use children's data for training helps neutralize immediate privacy backlash. But letting kids toggle the bot on by themselves—even with an email alert sent to parents—feels risky. An alert after the fact does not prevent an inappropriate interaction in the moment.

Offloading safety education onto parents looks like an easy out for Google. Teaching an 11-year-old to spot hallucinations or resist emotional attachment to a polite conversational bot is hard work. If Gemini slips up or offers questionable advice, Google will quickly find that warning emails are not enough to shield it from regulatory and parental fury.