Government IDs for age assurance: from selfies to zero-knowledge proofs (educatedguesswork.org)

The post analyzes how government IDs are being repurposed for legally mandated age assurance and why physical ID practices break down when moved online. In-person checks rely on tamper-resistant cards (holograms, UV features) plus a visual face match; remote options range from emailed scans (trivially spoofable) to selfies and live video. Video is stronger because it enables liveness checks, but attackers can still mount presentation attacks (fake cards, masks) or injection attacks (controlling the camera feed). Web contexts are particularly weak because browsers lack reliable software attestation and standards like C2PA are unlikely to secure live video. The modern technical response is cryptographic: issue digitally signed credentials (issuer signs DOB/photo) and bind them to a subject via asymmetric keys. The verifier checks the signature and uses a challenge/response (device signs a nonce with its private key) to prevent replay; photo matching still ties the credential to a person. This design prevents simple cloning of a public credential but remains vulnerable to credential sharing if keys are handed over, which is why many schemes add device binding. Crucially, digital approaches also enable privacy-preserving options—selective disclosure or zero-knowledge proofs to prove “over X” without revealing identity—offering a practical tradeoff between security, usability, and minimizing data exposure for age-verification systems.