Four questions decide whether data is an asset or a liability.
Buyers in this market converge on the same test, and most suppliers fail it quietly — in aggregate assurances, in licences written for distribution rather than training, in rights holders who were never actually paid. We answer all four in writing, per asset, before anything ships.
- 01
Are rights documented per asset, or claimed in aggregate?
Per asset. Every manifest row carries a consent_doc_ref and a contributor_licence_ref resolving to a scanned, signed document. Provenance travels attached to the file, not held in an inbox.
- 02
Does the licence explicitly name AI and machine-learning training?
Named in the grant. Our contributor licence grants the right to reproduce and process the material to train, pre-train, fine-tune, validate, benchmark and evaluate machine-learning and artificial-intelligence models, including generative models — and to sublicense that right onward through multiple tiers, so it passes to a laboratory without renegotiation.
Licences written for distribution, display or stock use are not accepted as a basis for supply, including our own contributors' prior agreements.
- 03
Can you produce the signed agreement for any given asset, on request?
Yes — name the asset ID. Consent is captured twice: a signed paper instrument, and a verbal on-camera confirmation recorded in the contributor's own language against the same asset ID. Either one alone would satisfy the test. We hold both.
- 04
Are the underlying rights holders actually being paid?
On the day, with a record. Contributors are paid at the point of capture and every payment carries a reference in a payment log tied to the assets it covers. Where a licence carries a revenue share, it is stated in the instrument rather than promised verbally.
- Appearance, voice & likeness
- Every identifiable person signs a plain-language release covering image, voice and voice likeness, including synthesis. Unreleased bystanders are blurred or the clip is cut, and the decision is logged in the manifest.
- Data protection
- Processing is on a consent basis under the Nigeria Data Protection Act 2023. Our informed-consent language states plainly that a model already trained cannot be untrained — so consent is given knowing what it means.
- Biometric exposure
- Faces and voiceprints are treated as biometric data. Releases are structured to withstand the standards buyers apply downstream, including Illinois BIPA and Texas CUBI verification.
Name any asset ID in a manifest.
We will send you the signed agreement for it. That is the whole test, and it is the one most suppliers cannot pass.