Instagram will begin scanning messages sent to and from under-18s to protect them from “inappropriate images”, Meta has announced. The feature, expected later this year, will work even on encrypted messages, suggesting the platform plans to implement client-side scanning for the first time.
However, the update will not report inappropriate messages back to Instagram servers. Instead, only the user’s personal device will know if a message has been filtered, leading to criticism that the company is “grading its own homework”.
Meta said the feature is designed to help protect teens from seeing unwanted and potentially inappropriate images from people they are already connected to, and to discourage them from sending such images themselves. The broad description is similar to Apple’s “communication safety” setting, which blurs nude photos sent to children’s devices.
The move comes as Meta faces a lawsuit in New Mexico accusing it of failing to protect children. According to an unsealed legal filing, Meta estimates about 100,000 children using Facebook and Instagram receive online sexual harassment each day. Mark Zuckerberg is set to appear before US Congress on Wednesday for a hearing on child safety.
Alongside the scanning promise, Instagram announced immediate updates: under-19s will default to privacy settings that prevent anyone they do not follow from sending direct messages, and parents will be prompted to approve or deny attempts by children under 16 to loosen safety settings.
Arturo Béjar, a former Meta engineer, said the changes need to be accompanied by regular reporting on unwanted advances. “This is another ‘we will grade our own homework’ promise,” he said. “Until they start quarterly reporting on unwanted advances, how are we to know they kept their promise?”
