Age verification explained

Age verification is the process of confirming that a person is old enough for something — a purchase, a platform, a place, a service. Three terms orbit it and are worth fixing precisely. Age verification checks a claim against evidence. Age estimation infers age from characteristics such as a face, with uncertainty attached. Age assurance is the umbrella for both — any means of reaching confidence about age.

The reason this subject has become technical is that the conventional method — show your identification, reveal your exact birth date — over-answers the question. A yes/no question deserves a yes/no answer, and the gap between the two is where privacy is lost.

Five approaches cover most age checks people encounter. Each has a genuine strength and a genuine cost:

The last row is the one that fits the question being asked. A zero-knowledge proof answers “over 18: yes” and nothing else — no date of birth, no document number, no photograph of a card.

Showing ID works, which is why it spread — but the working is done by over-disclosure. The service learns your exact birth date, your address, your document number and whatever else the card carries, and then decides what to keep. Most keep everything, because deletion takes a decision and retention takes none.

That accumulated material is the honeypot problem in miniature: an age check should not manufacture a dossier. There is also an arms-race problem — document images are cheap to fabricate now, so the check a service trusts may be weaker than it looks. The synthesis argument is set out in AI-generated fraud and fake IDs.

Age checks sit inside the wider principle of proportionality: the intrusiveness of a check should match the risk of the action it gates. A low-stakes age confirmation does not justify document-grade evidence; a legally sensitive transaction might. Applying this consistently is what verification tiers and proportionality argues for, and it cuts both ways — a service demanding maximal evidence for minimal risk is over-asking.

Proportionality is also an inclusion principle. Every rung of assurance that the risk does not require is a hurdle for someone who lacks documents, hardware or stable circumstances. Designing age checks around thresholds instead of documents lowers that hurdle without lowering the bar.

The Identity Wallet verifies you once — a document check plus a selfie under active liveness, at the Standard level — and anchors a commitment to the result. After that, an age question is answered with a proof generated on your device from your verified date of birth, without revealing the date itself.

Services receive a proof they can verify mathematically: no document, no date, no image. Your side of the experience — approving the request, seeing who asked and why — is covered in proving your age without your documents, and what each credential reveals is catalogued at credentials and proofs.

Age assurance is only ever as strong as the verification underneath it. A flawless threshold proof still rests on the enrolment that produced the verified birth date, which is why the enrolment step deserves the scrutiny. Estimation methods, meanwhile, are honest about uncertainty and should never be presented as identity checks — they say something about a face, not about a person's records.

No method is universally available: someone always lacks the document, the device or the account a given approach assumes. A fair system offers a route for each, and says plainly where the gaps are rather than pretending they do not exist.