Article 50(2) Marking-Conformance Testing: Proving Marking Survives Your Pipeline
By the Ingenious AI team · Published: November 14, 2026 · Reviewed: November 14, 2026 · 8 min read
The short answer
Register the system in the Evidence Vault, upload a sample output, and run the battery: re-encode, resize to half, format conversion and metadata strip, executed on the real bytes, with marking re-inspected after each transform. The run is hash-chained and signed into the system's history. Free covers one sample battery per system; repeated runs and pack export are paid.
A policy that says outputs are marked is an intention; a battery run on this version's actual output bytes is a measurement, and only the measurement survives an audit.
The Evidence Vault answers a provider-side question: does this generative system's output stay machine-readably marked after a realistic pipeline touches it? A system is registered in the vault, a sample output is uploaded, and the conformance battery runs real transforms on the real bytes: re-encoding, resizing to half size, format conversion, and metadata stripping. After each transform the bytes are re-inspected for marking, C2PA and JUMBF markers, XMP, a ProvenanceGuard assertion, or AI disclosure in metadata, and the before-and-after results are recorded as measured, never assumed.
What is marking-conformance testing? measuring, on real sample outputs, whether machine-readable marking is present before a pipeline's transforms and still detectable after each one, with the results sealed into a verifiable run history.
Why Article 50(2) conformance matters
Each battery run is sealed into a hash chain: the run's pack hash covers the previous run's pack hash plus the canonical record of this run, and the pack is HMAC-signed. Tampering with any stored field breaks recomputation, and the chain can be checked through the vault's API. That structure is what turns a folder of test results into evidence: a provider preparing for the Article 50(2) marking duties can show not just a passing run, but an unbroken, signed history of runs across versions. The free tier includes one sample battery per system; conformance history across repeated runs is a Pro and Team feature, as is evidence-pack export.
The legal backdrop in the product's own evidence wording is specific. The EU AI Act's Article 50 transparency duties have applied since 2 August 2026, and the marking grace period for systems already on the market before that date ends 2 December 2026. California's AI Transparency Act, SB 942, is operative from 2 August 2026. New York's synthetic performer law, in GBL section 396-b and effective 9 June 2026, covers disclosure where commercial advertising features a synthetic performer. The evidence packs carry disclosure wording for exactly these three jurisdictions.
How the record is built
None of that makes a provenance record a legal verdict, and ProvenanceGuard's packs say so themselves: evidence aid, not legal advice. What a record does is give counsel, a compliance team or a regulator something concrete to work from: the file, its hashes, the tool that sealed it, the timestamps, the declared disclosure and the wording used. Organisations that can produce that pack in minutes answer questions very differently from organisations reconstructing a publication history from chat logs and memory.
The practical reading of these dates is about preparation rather than panic. Duties that have already applied cannot be met retroactively for assets published unmarked, but a workflow adopted now, applied consistently to everything published from here on, with records kept per asset, is a defensible position that improves with every week it runs. The grace deadline in December 2026 is specifically about marking for pre-existing systems, which makes conformance testing of the marking itself, not just the policy around it, the urgent piece of work.
Where it meets the real world
Inspection comes before sealing, because an asset may already carry signals. ProvenanceGuard reads the actual bytes of an uploaded JPEG or PNG and reports what is really there: C2PA and JUMBF markers, an XMP packet, EXIF data, PNG text chunks, any AI disclosure already present in the metadata, and any prior ProvenanceGuard assertion, including whether that assertion's signature still verifies. Nothing is inferred from the filename or the uploader's claims; every field on the inspection result is derived from the file itself.
That byte-level honesty changes decisions. A file that arrives with a valid C2PA manifest needs different handling from a file whose metadata was already stripped by an earlier platform hop, and both differ from a file that was never marked at all. Checking first stops a common mistake: sealing and republishing an image while assuming its original manifest is intact, when in fact the copy in hand lost it two uploads ago. The inspection result becomes the baseline the rest of the record is built on.
Putting it into practice
Inspection also has limits worth stating. A marker being present says the bytes contain it; it does not say the claims inside it are true, and a signature check says whether a ProvenanceGuard assertion was altered, not whether the disclosure text a creator typed was accurate. Provenance records what was declared and when it was registered. Treating it as more than that, in either direction, is how teams end up both over-trusting and under-using the records they have.
Only JPEG and PNG files are supported, up to 25 MB, because those are the formats whose byte structures the inspection, embedding and fingerprinting are built and tested against. A file in another format is refused rather than guessed at. That refusal is a feature: a provenance tool that silently approximated on formats it does not understand would produce records that look solid and are not, which is worse than no record at all.
Five actions to take
- Run the battery per system version, not once per company
- Keep the chain head hash with your compliance file so history is checkable
- Seal the final export of the asset, with the creator and disclosure declared in your own accurate words
- Download and keep both copies the seal produces: the sealed copy as the record, the labelled copy for publication
- File the exhibit's verification URL beside the asset wherever your team keeps campaign or client records
A worked example: A small generative-image provider preparing for the December 2026 grace deadline
It registered 2 system versions in the vault and ran batteries on both: the older version's marking vanished after the re-encode transform, the newer one retained XMP marking through 3 of 4 transforms, and the signed chain showed exactly when the fix landed.
Notice what made the difference: not a better argument, but a record created before it was needed, carrying dates and hashes a stranger could check. That is the whole pattern this guide recommends, applied at the scale of one team and one deadline.
FAQ
Does a passing battery make a system compliant?
No. It evidences measured marking survival, which is one input to a compliance position the provider must still take with counsel. The vault records measurements and states that limit in its own pack.
Is ProvenanceGuard's assertion a C2PA manifest?
No. It is ProvenanceGuard's own signed assertion, embedded as XMP and HMAC-signed, and labelled that way everywhere. The inspection detects genuine C2PA markers when a file carries them, but the two records are never conflated.
What formats and sizes are supported?
JPEG and PNG, up to 25 MB per file. Those are the formats the inspection, XMP embedding and fingerprinting are built and tested against; other formats are refused rather than approximated.
What does the free tier actually do?
The complete job on an asset: byte-level inspection, a signed assertion embedded as XMP, a visibly labelled copy, SHA-256 and perceptual fingerprints registered, and a public verification page. Paid plans add survival re-checks, evidence packs and bulk verification.