Two US senators have asked Meta how smart-glasses images are accessed and used for AI training.

The big change

  • Recording and subsequent AI training are distinct stages of the inquiry.
  • The senators ask whether people captured can request deletion.
  • Meta described its capture light and storage practices in July.

What the senators want to know

Senators Ron Wyden and Edward Markey sent their letter to Meta CEO Mark Zuckerberg on October 28; Markey’s office announced it October 29. They requested answers by November 25.

Citing reports of contractors reviewing intimate footage, they ask about transmission, human access, filtering and use in training, alongside indicator testing and deletion rights for owners and bystanders. Their questions do not establish that any particular bystander’s image trained a model.

Meta’s earlier explanation

In a July explanation, Meta said photos and videos taken for a wearer’s gallery are stored on the glasses until imported to a phone. It said a front-facing light blinks during gallery capture, and that from the second generation the camera is disabled if the system detects the light is blocked.

The company also said it was updating the glasses to disable the camera when physical tampering with the light is detected. Those are Meta’s descriptions, not an independent test.

The earlier post does not address the senators’ questions about contractor access or training.

BIG CHANGE found no public Meta response to the letter in the sources reviewed.

BIG CHANGE previously covered European data protection guidance on smart-glasses privacy signals, a separate development.

Sources & further reading