Re:atproto and the ownership of identity
by lxgr on Sunday January 19, 2025 @02:16PM
(#42760566)
"Proof of <x>hood" is explicitly a non-goal for FIDO authenticators, and implementors go to great lengths to make it impossible (primarily for privacy reasons); that's why they support anonymous attestation only, except in an enterprise context.
That's the thing about reusing existing services and protocols for "proof of x": Quite often, such proofs being possible are often adjacent to or explicitly an anti-goal.
Apple does indeed offer an API that allows uniquely identifying a given device in a privacy-preserving way, though: There's an API that allows apps to store two bits of information about a given device, such as "this device was involved in fraudulent activity on our platform" or "this device has already redeemed a free trial at one point". Not sure if Google has an equivalent.
That's the thing about reusing existing services and protocols for "proof of x": Quite often, such proofs being possible are often adjacent to or explicitly an anti-goal.
Apple does indeed offer an API that allows uniquely identifying a given device in a privacy-preserving way, though: There's an API that allows apps to store two bits of information about a given device, such as "this device was involved in fraudulent activity on our platform" or "this device has already redeemed a free trial at one point". Not sure if Google has an equivalent.