# Data provenance

URL: https://sellmydatatoai.net/glossary/data-provenance

Data provenance is the documented record of where a dataset came from, who created and owned it, and how it was collected, changed and transferred.

Data provenance is the documented history of a dataset: its original source, who created and owned it, how it was collected, what was changed and every transfer since. For AI buyers it is the evidence that the data can lawfully be used for training.

## What provenance evidence do AI buyers expect?

AI buyers expect provenance evidence for each asset, not just a license tag. Getphyllo lists six things a training data license should cover:

- Rights documentation for each asset
- Explicit permission for AI training
- A documented chain of acquisition
- Clearance for each layer, such as music, images and performances inside video
- Evidence that rights holders were paid
- Withdrawal terms, including for models already trained

## Does the law require data provenance?

The law requires some provenance disclosure in California. AB 2013 requires developers of generative AI systems to post a high-level summary of the datasets used in development on their websites, on or before January 1, 2026, according to a [California Attorney General legal advisory](https://oag.ca.gov/system/files/attachments/press-docs/Legal%20Advisory%20-%20Application%20of%20Existing%20CA%20Laws%20to%20Artificial%20Intelligence.pdf). [NIST](https://csrc.nist.gov/glossary/term/provenance) defines provenance as the chronology of the origin, development, ownership, location and changes to a system and its associated data.

In the EU, [Article 53 of the EU AI Act](https://eur-lex.europa.eu/eli/reg/2024/1689/oj) requires providers of general-purpose AI models to publish a sufficiently detailed summary of training content, covered in [EU AI Act training data transparency](https://sellmydatatoai.net/guides/eu-ai-act-training-data-transparency). See [data provenance explained](https://sellmydatatoai.net/guides/data-provenance-explained) and [chain of title](https://sellmydatatoai.net/glossary/chain-of-title).