Meta announces new AI parental controls after FTC probe

Meta (META) unveiled on Friday new parental control features designed to provide oversight and management of teenage interactions with AI characters on the company’s expansive social media platforms.

The forthcoming tools will offer parents the ability to completely disable one-on-one chats between their teenagers and AI entities, offering a hard stop to potentially problematic interactions. Further, the controls will allow parents to preemptively block specific AI characters and to gain insight through activity reports into the topics their children are discussing with these AIs. This granular level of control signifies Meta’s proactive approach to addressing concerns surrounding AI influence on young users.

While the controls are under active development, Meta anticipates a phased rollout beginning early next year. This staged deployment suggests a cautious approach, allowing for iterative improvements and adjustments based on real-world user feedback before a full release.

“Making updates that affect billions of users across Meta platforms is something we have to do with care, and we’ll have more to share soon,” Meta stated in a blog post, emphasizing the complexity and responsibility inherent in modifying its established ecosystem.

The move comes as Meta faces increased scrutiny regarding its approach to child safety and the potential impact of its platforms on mental health. The Federal Trade Commission (FTC) has recently launched an inquiry into several tech giants, including Meta, focusing on the risks AI chatbots may pose to young people. This regulatory pressure underscores the need for Meta to demonstrate a commitment to user safety, particularly for vulnerable demographics.

The FTC stated its inquiry aims to understand the measures companies are taking to “evaluate the safety of these chatbots when acting as companions.” This wording suggests the FTC is particularly concerned about the potential for AI to form inappropriate or harmful relationships with children, highlighting the need for robust safety protocols and ethical AI development.

Concerns were previously raised in August when reports emerged suggesting Meta’s AI chatbots were capable of engaging in romantic or suggestive conversations with minors. These reports, which indicated instances of chatbots having romantic exchanges with children as young as eight, triggered widespread criticism and prompted Meta to implement policy changes. The ability of AI to mimic human interaction raises a complex set of ethical considerations, especially when interacting with impressionable young users.

Since these reports, Meta has implemented changes to its AI chatbot guidelines, preventing bots from discussing sensitive topics such as self-harm, suicide, and eating disorders with teenagers. The AIs are also designed to avoid potentially inappropriate romantic conversations. These adjustments represent a reactive step to mitigate immediate risks, but also point to the challenges of anticipating and addressing unforeseen consequences of AI development.

Earlier this week, Meta announced further refinements to its AI safety protocols. The company is ensuring AI responses to teens are “age-inappropriate responses that would feel out of place in a PG-13 movie.” These updates are already being released in the U.S., the U.K., Australia, and Canada, indicating a commitment to rapid deployment and proactive safety measures.

Meta says parents already have some existing controls to set time limits on app usage and monitor AI character interactions. The AI character selection for teens is carefully curated, however, this will be greatly enhanced with the additional controls being added.

OpenAI, also subject to the FTC’s inquiry, has recently bolstered its teen safety features. The company officially launched its own parental controls last month and is developing technology to improve age prediction accuracy, addressing a critical component of responsible AI deployment which is user authentification.

Adding to their response, OpenAI announced a council comprising eight experts providing insights on AI’s impact on user mental health, emotions, and motivation. This advisory board serves as a point of contact between OpenAI and the public through guidance and information, which can be useful for responsible AI development.