Google makes Gemini chatbot accessible for children under 13 – via Family Link
Parents can use Family Link to control how their children use Google services. They were informed by email that their children can also use a chatbot.
Google Gemini-App.
(Image: mundissima/Shutterstock.com)
Children of parents or guardians who use Family Link to control their children's use of Google services now have access to Google's Gemini AI chatbot. Actually, only people who indicate that they are over 13 years old can use Gemini. And Family Link is actually intended to allow parents to restrict their children's Google accounts in general, for example YouTube consumption.
Google's current move is more evidence of the company's reverse understanding: accounts of under 13-year-olds linked to Family Link are subject to special supervision –, so why not make the chatbot accessible to precisely these children, while those without parental Family Link support are left out? According to the New York Times, the first parents have now apparently been informed by email that Google has put this idea into practice.
Chatbot for children from next week
According to the report, the Gemini chatbot will be available next week (from May 5) for Google accounts of under 13-year-olds who meet the above-mentioned requirements. The parent of an eight-year-old child was told in an email that they will soon be able to use Gemini "to ask questions, get help with homework and make up stories." So far, all media reports on this have come from the US; it is unclear whether German users have also been affected.
A German-language Google support page states that Google wants to introduce Gemini "gradually in various markets for accounts with parental control". This means that children will be able to use the Gemini app and web view, as well as access Gemini via Google Assistant. Google believes that a responsible introduction to Gemini lies with the parents themselves: "Talk to your child about the fact that Gemini is not human," is one of the general pieces of advice on the support page. Alternatively, according to Google, parents can also deactivate their child's access to Gemini apps via the following link in the settings. According to the company, the minimum age in Germany for a Google account with full functional access is 16. So it could well be that the sudden Gemini access via Family Link here will apply to all children under the age of 16.
"Strict safety guidelines" for children
Meanwhile, a Google spokesperson assured the New York Times that Gemini has particularly strict safety guidelines for children so that the chatbot avoids certain harmful content. However, such safety guidelines are usually easy to override, probably even for one or two 13-year-olds. Recently, for example, heise online tested a trick that common chatbots such as Gemini use to give advice on things that go well beyond the proverbial "silly boy pranks", such as making Molotov cocktails. The applications are actually supposed to refuse to provide such information.
Videos by heise
Parents also receive mail on first use
The Google spokesperson also told the New York Times that the data from the children's Gemini conversations would not be used to train the AI model behind the chatbot. Google also confirmed to tech magazine TheVerge that parents would also receive an email when one of their children uses the chatbot for the first time.
The New York Times also sees Google's move as an attempt to establish its AI applications among the youngest children and talks of a real competition between tech companies, for example at schools, universities or in leisure activities. The sudden chatbot access for children is not entirely new: many parents have probably already noticed that their children can chat with the Meta AI chatbot via WhatsApp, Facebook and Instagram. There are also known cases of sexualized conversations with minors. In general, when it comes to AI applications, something can now also be observed among users who are particularly in need of protection, which was first seen in a completely different place with the forced use of Copilot in Microsoft365 subscriptions: it is not about offering users AI functions and convincing them to use them – but about making them as unavoidable as possible.
(nen)
