Decisions · Policy · United States

Generative AI meets copyright

Pamela Samuelson (UC Berkeley School of Law) · 2023-07-14 · Proposal · non-binding

Science, vol. 381, issue 6654, pp. 158 to 161 (2023)

AI training and copyright

Can AI be trained on protected works?

Closer to freely allowed

If the lawsuits succeed, lawful US generative AI could be limited to public-domain or licensed data

Why hereCloser to 'freely allowed': the article warns that plaintiff wins could confine lawful US systems to public-domain or licensed data, treating that outcome as a cost.

What it says

Surveying the first copyright lawsuits against generative AI, the article explains creators' objections to training and to competing outputs, and warns that plaintiff wins could confine lawful US systems to public-domain or licensed training data. In related talks Samuelson urged showing how generative AI can advance copyright's purposes.

A leading US voice warning that copyright litigation could reshape who can build AI.

Sources

This record

Data as of 2026-10-21.

Drafted with AI assistance and checked against the sources linked above; not yet reviewed by a lawyer. Methodology

No changes logged yet. The change log started on 2026-10-20.