What regulators are asking for
The EU AI Act asks providers of high-risk AI systems to apply data governance to training, validation and test data, including the origin of the data, how it was collected, and how it was prepared, such as labelling. General-purpose model providers are asked to publish a summary of the content used for training.
Other frameworks, such as the NIST AI Risk Management Framework, point the same way: know your data and be able to show it.
What comes with our data
Every delivery carries records your compliance team can use:
- Consent: a reference to the signed agreement or consent behind each item
- Provenance: who produced each item and who reviewed it, and when
- Preparation: the spec, labelling guidelines and QA results for each batch
- Licence: the terms you hold for the data, set in the order form
What this is not
This page is not legal advice, and using Saolabs data does not by itself make a model compliant. Your obligations depend on your system and how it is used. Talk to your counsel about what applies to you.