Instagram will let parents block their kids from talking to 'AI characters'

Facebook and Instagram parent Meta is introducing a new feature to let parents control how — and if — their teen can interact with AI characters. 

The new update will let parents block some AI characters or shut off access altogether for their teen’s account. But even if parents opt to block all access to AI characters, their teen will still have access to Meta’s AI assistant. Meta said this feature will have “default, age-appropriate protections in place.” The update will first be rolled out on Instagram early next year, the tech giant said Friday. 

Meta's latest update to teen protections comes just months after a Reuters report found that an internal policy document at the company let its AI chatbots engage in romantic conversations with kids. 

In response to this report, attorneys general for 42 states and two territories sent a letter to 13 companies using artificial intelligence — including Meta — saying they would use “every facet of our authority to protect children from exploitation by predatory artificial intelligence products.”

Meta isn’t the only company playing catch up on teen protections. OpenAI, the company behind ChatGPT, rolled out a slew of parental controls in September, including a teen version of ChatGPT.  

In addition to Friday’s update, Meta announced on Tuesday that it updated “AI experiences” for teen users to follow guidance of PG-13 ratings. It said this means “AIs should not give age-inappropriate responses that would feel out of place in a PG-13 movie.” 

“In addition to our longstanding policies — which already hide or prohibit the recommendation of sexually suggestive content, graphic or disturbing images, and adult content like tobacco or alcohol sales from teens — our updated policies will now go even further,” Meta said. 

Parents will have the option to add even stricter settings to their child’s account, which would filter “even more content” than the PG-13 version. The update is being rolled out in the U.S., U.K., Canada, and Australia in English. Meta said its AI characters already “are designed” to not discuss self-harm, suicide, or eating disorders with teens. 

Meta said parents can get access to “insights” into what topics their child is talking about with AI characters and the AI assistant. The tech company did not elaborate more specifically on what these insights would include. 

Users who Meta believes could be teenagers will have these age-restrictions placed on their accounts. However, researchers have documented how easy it is to circumvent limits set by chatbot companies. Age-verification rules are also known to be easily bypassed.