873
Group Pushing Age Verification Requirements for AI Turns Out to Be Sneakily Backed by OpenAI (gizmodo.com)
submitted 7 months ago by to c/technology@lemmy.world
46 comments fedilink hide all child comments
๐Ÿ˜ก 1
you are viewing a single comment's thread
view the rest of the comments
[โ€“] 20 points 7 months ago (2 children)

Matthew Green the CS/cryptography professor is actively writing about this in fairly broad language https://blog.cryptographyengineering.com/2026/03/02/anonymous-credentials-an-illustrated-primer/

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [โ€“] 8 points 7 months ago (1 child)

    tl;dr: The "zero knowledge" proof could have a finite number of uses per block of time for each verifier, each of which represented by a unique single-use key. This way anyone sharing keys would be limited by that finite number of uses, and if people sharing this aren't coordinated they could end up re-using a single-use key.

    If the encryption was stolen without their consent, this could tip a user off prompting them to invalidate the current set and get a new one. And if the verification is used to support a pseudonym like an account for an online service then instances of re-use could get flagged for moderators.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •