Science & Tech · AI · 4 hrs ago
Scholars question whether AI competition can support copyright claims
Three copyright scholars say claims of AI market harm need stronger proof.
They argue that AI-made books and images can compete with human creators, but competition alone has not traditionally counted as copyright infringement.
The debate grew after a US judge ruled that Meta’s use of books to train its Llama models was fair use, while warning that evidence of market harm could have changed the outcome.
The scholars say “market dilution” has no clear definition, and research on AI’s effects on creators points in different directions.
One stock-photo marketplace saw AI-image uploads rise 136% as uploads of non-AI photos fell 15%, but those figures alone do not settle whether human creators were harmed.
More than 100 US lawsuits are testing AI copyright questions, with consequences for creators and the companies that build and use AI systems.
The scholars say courts should require evidence linked to a specific AI system and market before accepting dilution claims.
Three copyright scholars argue that claims of AI-driven market dilution should require evidence tied to a specific AI system and market.
They say the concept has no clear definition and existing studies conflict on whether AI harms human creators.
The debate follows a 2025 ruling that found Meta’s use of authors’ books to train its Llama models was fair use.
The judge said evidence that AI outputs weakened authors’ incentive to create could have changed the ruling.
More than 100 US lawsuits are testing whether training AI on copyrighted books and images is legal.
- Who
- Copyright scholars Pamela Samuelson, Abhishek Nagaraj and Christopher Jon Sprigman question market dilution claims.
- What
- They argue that courts should require evidence tied to a specific AI system and market before accepting such claims.
- When
- The essay was reported on October 16, 2026; the Meta ruling was in June 2025.
- Where
- The debate concerns copyright lawsuits in the United States.
- Why
- The scholars say market dilution lacks a clear definition and research on AI’s effects on human creators is conflicting.
The three copyright scholars
The federal judge who ruled in the Meta case
Need for proof
The three copyright scholars
Courts should not accept market-dilution claims without evidence tied to a specific AI system and market.
The federal judge who ruled in the Meta case
The judge said evidence that AI market harm occurred could have defeated Meta’s fair-use defense.
Meaning of market harm
The three copyright scholars
Competition from AI outputs alone should not count as copyright infringement when no protected expression is copied.
The federal judge who ruled in the Meta case
The judge warned that large volumes of cheap AI works could weaken humans’ incentive to create.
A US Copyright Office report briefly mentioned market dilution, according to the scholars.
A judge ruled that Meta’s training of its Llama models on authors’ books was fair use, while warning about possible market harm.
Earth.com reported on the scholars’ arguments against accepting market-dilution claims without specific evidence.
- Scholars
- Pamela Samuelson, Abhishek Nagaraj and Christopher Jon Sprigman
- Lawsuits
- More than 100 in the United States are testing whether AI training on copyrighted books and images is legal
- Meta case
- Richard Kadrey and 12 other authors sued Meta over training its Llama models on their books
- Ruling
- June 2025; the judge ruled for Meta on fair use
- Stock-photo uploads
- AI-made image uploads rose 136% after a large marketplace began allowing them
- Non-AI photos
- Uploads fell 15%, and the number of artists posting them fell 29%
What people are saying
market dilution