Two Harvard students have created an eerie demo of how smart glasses can use facial recognition tech to instantly dox people’s identities, phone numbers, and addresses. The most unsettling part is the demo uses current, widely available technology like the Ray-Ban Meta smart glasses and public databases.

AnhPhu Nguyen, one of the two students, posted a video showcasing the tech in action that was then picked up by 404 Media. Dubbed I-XRAY, the tech works by using the Meta smart glasses’ ability to livestream video to Instagram. A computer program then monitors that stream and uses AI to identify faces. Those photos are then fed into public databases to find names, addresses, phone numbers, and even relatives. That information is then fed back through a phone app.

In the demo, you can see Nguyen and Caine Ardayfio, the other student behind the project, use the glasses to identify several classmates, their addresses, and names of relatives in real time. Perhaps more chilling, Nguyen and Ardayfio are also shown chatting up complete strangers on public transit, pretending as if they know them based on information gleaned from the tech…

  • sp3ctr4l
    link
    fedilink
    arrow-up
    15
    ·
    3 months ago

    Potentially fashionable solution:

    Get your own glasses, get some sufficiently powerdul IR emitting LEDs, attach them to the frame, wire them to a battery.

    Result: Most cameras will see your face as a blinding white spot.

    Apparently ‘Reflectables’ claims to achieve a similar result passively, with reflective frames and lenses.

    • EldritchFeminity@lemmy.blahaj.zone
      link
      fedilink
      arrow-up
      7
      ·
      3 months ago

      I’ve heard of coats with that passively blinding IR effect before, though the active one is basically IRCM or infrared dazzlers. Another option would be asymmetric makeup with sharp geometric shapes - anything that would help break up the normal identifiers of a face. I remember seeing this proposed many years ago as a countermeasure to facial recognition software.