Sign in to your Android Police account
Summary
- Google is launching a kid-friendly version of its Gemini AI chatbot for children under 13, available only through Family Link.
- While Google says it won’t use kids’ data to train Gemini and has added content filters, it admits the filters aren’t perfect.
- Parents will have the option to disable Gemini entirely if they choose.
A couple of days ago, ChatGPT was caught allowing minors to generate erotica, as reported by TechCrunch. The AI bot even encouraged a few of these minor users to ask for more explicit content. OpenAI was quick to acknowledge the issue and told TechCrunch that they’re “actively deploying a fix,” noting that their policies don’t allow such responses to be shown to users under the age of eighteen.
Related
6 things I had no idea Gemini could do
Google Gemini just got even more useful for me
While it was confirmed that the issue was due to a bug, it goes to show just how dangerous AI can be, especially in the hands of youngsters. While users need to be above the age of thirteen to use both OpenAI and Google Gemini, Google has plans up its sleeve to roll out Gemini to children under 13 as well. Unfortunately, although Google is directing its efforts toward ensuring a safe experience for younger users, part of me is afraid it won’t end up all too well.
A kid-friendly version of Gemini is launching next week, but it can be disabled
As reported by The New York Times, Google is planning to roll out its Gemini AI chatbot for children under the age of 13. The kid-friendly version of Gemini will only be available to children under 13 who use Family Link, which allows parents or guardians to set up a Google account for their child. Family Link also lets parents supervise their child’s account by deciding which apps their child can install, blocking certain apps, changing app permissions, setting screen time limits, monitoring usage, and even restricting mature content on Google Play.
Google has started emailing parents who use Family Link controls to manage their child’s device, mentioning that Gemini apps will soon be available for their child. This means they’ll be able to use the AI assistant to “ask questions, get homework help, and make up stories.”
Source: Google
While Google will not use the data of a child with a Family Link account who uses the kid-friendly Gemini to train its AI, the tech giant acknowledged the risks of AI in the email. Google clarified that Gemini can indeed make mistakes, and suggested parents “help your child think critically about Gemini responses.” They also recommended reminding children that Gemini isn’t human, showing them how to double-check its responses, and avoiding the sharing of sensitive or personal information when using Gemini.
Google noted that its filters “try to limit access to inappropriate content, but they’re not perfect.” Ultimately, this could result in children seeing content that parents might find inappropriate.
The good news? If you’d rather your child not use Gemini at all, even the kid-friendly version, Google spokesperson Karl Ryan confirmed to The Verge that parents will have the option to disable it via Family Link. He also mentioned that parents or guardians will be notified when their child uses Gemini for the first time.
