Do you have the rights to license this dataset? A four-layer checklist
In the United States, raw facts are not copyrightable, so the rights in a dataset usually come from somewhere else: contracts, trade secret status, and the copyright in the works it contains. Privacy law then limits what any of those rights let you do.
Where the right to license comes from
- Copyright Included works and original compilation.
- Contract & trade secret Customer, vendor and platform terms.
- Privacy limits Who is identifiable and what they were told.
Original source
Who created or collected it?
Rights chain
What contracts and permissions apply?
Proposed AI use
Does the allowed scope cover this use?
Check copyright, contract restrictions and personal-information obligations separately. Access alone is not permission to license.
The layers
| Layer | Source of the right | Question to answer |
|---|---|---|
| Copyright in included works | Copyright Act; authors or assignees | Does the set contain articles, images, recordings or code someone else authored? |
| Compilation copyright | Original selection/arrangement (Feist) | Is your structure creative, or merely exhaustive and alphabetical? |
| Contract | Your terms with customers, vendors, platforms | Does any agreement limit reuse, transfer or AI use? |
| Trade secret | DTSA and state law | Would licensing disclose information you protect as secret? |
| Privacy (limit, not a right) | FTC Act, CCPA, HIPAA, BIPA, COPPA | Is anyone identifiable, and what were they told? |
No 'database right' in the US
Unlike some jurisdictions, the US has no standalone right protecting databases for the investment that went into them. After Feist, your leverage over a factual dataset rests mostly on access control and contract. That is why license terms and delivery method matter so much.
Fair use doesn't settle your position
Whether training on unlicensed material is fair use is being argued in courts and was analyzed in the Copyright Office's report on generative AI training. Either way, it doesn't give you, the seller, rights over works you didn't author, and buyers paying for a license typically want clean title regardless.
How this affects value
Clean title is what a license fee pays for; each unresolved layer narrows the grant you can offer.
See what shapes a dataset's value