What the unit asks
The working draft adopts the ISO/IEC 29155 benchmarking model: the thing being compared is how a system performs, not the company that sells it. That distinction only holds if the boundary is written down. A face-estimation component tested on a bench is not the same object as the same component inside a document-checking workflow, and the two results cannot be pooled.
What good practice looks like
Write one sentence per object, name the component chain and mark what is excluded. Version the sentence. When a supplier changes a model, a threshold or a document list, that is a new object and a new comparison.
Evidence an assessor asks for
| Artefact | Why it matters |
|---|---|
| An object register with one row per system or component, its version, its supplier and the date the version was confirmed. | It is the only defence against comparing unlike things and calling the difference a result. |
| A boundary diagram showing the components, the processing sub-system and the relying party interface. | It shows what the test covers and what it leaves outside. |
| A change log for the comparison period. | A model or threshold change mid-test invalidates the comparison, and the log is how you know. |
What a good answer sounds like
A good answer to P3-01 names An object register with one row per system or component, its version, its supplier and the date the version wa, dates it, and says what it does not cover. A policy or a promise is not an answer. If P3-01 has no artefact, the row is not met and the absence is the finding.
Where comparisons commonly fail
- The comparison mixes a component test with a system test and reports one number.
- The version of the system tested is not recorded, so the result cannot be tied to the service that is live now.
- The relying party interface is left outside the boundary, although it decides what data is disclosed.
Questions to put to the supplier
- What exactly is the object, and what is outside it?
- If the supplier ships a new model tomorrow, is this comparison still about the same thing?
How this is used in the kit
P3-01 becomes one row of the comparison register, with an owner, a status and the evidence named above. A status of met asks for the artefact and its date, and the free check uses the same row as an evidence question, so the page and the register share one vocabulary.
Source for this page
- ISO/IEC JTC 1/SC 27/WG 5 N3928, AWI 27566-2 working draft. Public working draft of the benchmarking part, at the ETSI document box.
- Measurement of Age Assurance Technologies. The DRCF and ICO research report on measuring age assurance technologies.
Retrieved 2026-10-21. P3-01 is our statement of WG 5 N3928, clause 4.1.1; the DIS body is not public, so no DIS clause number is claimed.