What the study found
The article argues that data is nothing like oil and, for copyright purposes, should not be treated as though it were. It says data is intangible, not finite, and neither rivalrous nor excludable, which makes it a public good; in copyright terms, the author suggests it should look more like the ocean than oil.
Why the authors say this matters
The authors say this matters because AI systems rely on large data sets that often include copyright-protected works, and current disputes, policy consultations, and private licensing deals are all testing how copyright should apply. The study suggests that the metaphors used to describe data may shape how people understand and regulate AI.
What the researchers tested
This is a research article by Carys J. Craig that examines the idea that “data is the new oil” and compares data with oil both economically and under copyright law. It discusses copyright infringement cases, government consultation papers, policymakers’ questions, and licensing arrangements involving AI training data.
What worked and what didn't
The article says the oil analogy works only in limited ways: both data and oil can create economic value, power, and environmental harm. It says the analogy breaks down because data is not tangible, finite, rivalrous, or excludable, and that AI training datasets often include copyrighted works that are now the subject of infringement disputes.
What to keep in mind
The abstract does not report empirical experiments or specific case outcomes. It also does not describe limitations beyond the scope of the argument, which is mainly conceptual and legal.
Key points
- The article argues that data should not be treated like oil, especially in copyright law.
- It describes data as intangible, not finite, and neither rivalrous nor excludable.
- The author says data is better understood as a public good, or in copyright terms, something more like the ocean.
- AI training datasets often include copyright-protected works, which are already raising infringement disputes.
- The abstract frames the piece as a conceptual and legal argument rather than an empirical study.
Disclosure
- Research title:
- Copyright originality doctrine may better fit AI data than oil
- Publication date:
- 2026-03-01
- OpenAlex record:
- View
Get the weekly research newsletter
Stay current with peer-reviewed research without reading academic papers — one filtered digest, every Friday.