In its frantic quest for new users, Google intends to launch a version of Gemini for children under 13. Young users will be able to ask questions, ask for help with their homework, or even invent stories. Parents who manage Google accounts have already started receiving emails from the web giant, announcing that Gemini apps will soon be available for their children.
AI from kindergarten
So obviously, put like that, the prospect of leaving a child alone with an AI is a little scary. In a recent blog post, UNICEF warned about the dangers of AI bots for very young children: confusion between human and machine, alteration of the perception of reality, orientation of opinions and biases of thought, emotional manipulation, vulnerability to misinformation, and exposure to inappropriate content.
Google knows it has to walk on eggshells. That's why safeguards are in place. First, access is limited to children registered in accounts managed by parents (Family Link). The latter must provide personal information such as the child's name and date of birth in order to create the young user's account.
This version of Gemini includes specific filters to prevent the generation of content deemed dangerous or inappropriate. Parents are notified when their child accesses Gemini for the first time, and they can manage or block their child's access to the chatbot at any time through Family Link settings. Finally, Google assures that children's interactions are not used to train AI models.
Google also acknowledges that Gemini "can make mistakes" and encourages parents to teach their children critical thinking skills and supervise the chatbot's use. This new service will initially only be offered in the United States.
Source: New York Times
0 Comments