This week, Google reportedly sent an email to parents to let them know that the Gemini AI chatbot will soon be available for children under 13 years old.

The New York Times cites an email that states the chatbot would be available starting next week for certain users. (Chrome Unboxed reported on the same email on April 29.) Google sent the email to parents who use the company’s Family Link service, which lets families set up parental controls for Google products like YouTube and Gmail. Only children who participate in Family Link would have access to Gemini, for now. The email reportedly told parents their children would be able to ask Gemini questions or assist with tasks like doing homework.

The move comes days after the nonprofit Common Sense Media declared that AI companions represent an “unacceptable risk” for people under 18. Common Sense Media worked with researchers from Stanford School of Medicine’s Brainstorm Lab for Mental Health Innovation, resulting in a report urging parents to stop underage users from accessing tools like Character.ai.

Character.ai is one of a growing number of services that let users create and interact with AI “characters.” As Common Sense Media wrote in its report, “These AI ‘friends’ actively participate in sexual conversations and roleplay, responding to teens’ questions or requests with graphic details.”

This type of roleplaying is distinct from AI chatbots like ChatGPT and Gemini, but it’s a blurry line. Just this week, Mashable reported on a bug that would have allowed kids to generate erotica with ChatGPT, and The Wall Street Journal exposed a similar bug with Meta AI. So, while AI chatbots like Gemini do have safeguards to protect young people, users are finding ways to get around these guardrails. It’s a fact of life on the internet that some rules are easily skirted. Just consider online pornography, which is illegal for people under 18, yet widely available with just a few clicks.

So, parents who want to keep their kids from using artificial intelligence are facing an uphill battle.

To make the debate even more complicated, President Donald Trump recently issued an executive order that would bring AI education into U.S. schools. The White House says the order will “promote AI literacy and proficiency of K-12 students.” Understanding AI’s abilities, risks, and limitations could be useful for children using it for schoolwork (especially considering its tendency to hallucinate).

In its email to parents, Google acknowledged these issues, urging parents to “help your child think critically” when using Gemini, according to The New York Times.