As a slightly pyromaniac firefighter, for Meta, protecting minors on its platforms also means protecting them from its own chatbots. Since September 2023, the parent company of Facebook and Instagram allows users to interact with conversational agents played by celebrities, personalized AI created by other Internet users and its assistant, Meta AI.
Although the company was criticized this summer for allowing its chatbots to have “sensual or romantic” conversations with children, it seeks to reassure them. On Friday, October 17, it announced new controls so parents can better manage their children’s experiences.
Better control for parents
Meta had already provided a summary of these changes earlier in the week with the launch of two new features for “teen accounts” on Instagram. One of these novelties is a parameter aligned with the American film classification PG-13 (not recommended for children under 13 years of age) to avoid exposing young people to content inappropriate for their age (bad language, incitement to harmful behavior, etc.).
As the California group previously noted, it has also updated its AI-based experiences for teens to align with this classification as well. The chatbots available on their platforms should no longer give inappropriate responses that would be considered inappropriate in a movie not recommended for children under 13, he explained without elaborating.
Meta, however, wants to go further with new monitoring tools for parents. To begin with, the company will offer them an overview of the topics discussed by their children during their exchanges with these conversational robots. Soon they will also be able to completely disable their children’s access to individual chats.
Parents who do not wish to completely disable this access will have the option to block some of the chatbots.
Additional protections
These protections are in addition to measures already in place to provide appropriate experiences for “teen accounts” when interacting with AI. With this objective, parents can now see if their child is chatting with them, but also set usage limits. Chatbots should also be careful when discussing sensitive topics (suicide, self-harm, eating disorders, etc.) with young people.
With these new tools, the company hopes to “reassure parents that their teens can take full advantage of all the benefits that AI has to offer, with the right safeguards and monitoring.” However, they will have to wait a little before being able to use them because they will not be deployed until early 2026, to start on Instagram. And these tools will be available first in English in the United States, United Kingdom, Canada and Australia.
Source: BFM TV
