Data rights & provenance

Provenance.

Consent, license scope, and origin recorded at collection.

See licensing

The principle

Rights travel with the data.

We collect from consented contributors under written agreements. Consent and licensing are in place before recording starts.

Each collection records origin, consent, and license scope. Those records are available before delivery.

What provenance means here

Consent, license, origin, and redaction.

Explicit consent

Contributor agreements define how collected material may be used before recording begins, with the applicable consent record included in dataset review.

Clear licensing

The license states the permitted AI-development use and engagement scope before production delivery.

Traceable origin

Managed collection programs record the source and collection path instead of relying on unknown third-party aggregation.

Documented redaction

Where personal or sensitive material is in scope, redaction requirements are defined in the dataset specification and QA record.

Questions

Rights & provenance

Where does the data come from?
We source at the origin, directly from consented contributors who record to a specification, rather than scraping the open web. That is what lets provenance travel with each sample.
What rights come with the data?
Rights are established with the contributor at the point of collection and scoped per engagement. Licensing terms are documented and confirmed in writing before delivery. See the licensing overview for how scope is set.
Can we review the provenance record?
Yes. Robotics data programs define contributor consent, provenance, and licensing records with the collection specification, and include the applicable documentation in buyer review.