This article was prepared with AI using the sources below. Read those sources for their original context.

Translated from the Dutch article. The original publication date is retained.

Reports about Sony, Warner and Anthropic raise a broader question: which rights are needed for AI training and for use of the outputs? This article makes no finding about the merits or outcome of a particular lawsuit.

The U.S. Copyright Office addresses AI training and copyright in a separate report. A specific assessment matters; US rules or a case about one dataset do not provide general permission for use in the Netherlands.

What does this mean for your business?

Distinguish rights in your input material, a supplier’s use of that material and the outputs you publish. Those are different questions. Have uncertain cases assessed against the applicable rules and contracts.

A familiar supplier does not guarantee that every use is permitted under copyright law. Check the terms, rights to inputs and intended use of outputs. Have uncertain cases assessed.

What can you do tomorrow?

  • Read your AI supplier's terms. Do they say anything about copyright?
  • Do not use AI to generate text that closely resembles existing work, such as song lyrics or book extracts.
  • Ask your supplier whether they have arranged rights to training data. If not, consider another provider.

Do not wait for a general ruling that would resolve every use. Record permitted inputs and outputs for each application and the checks before publication. That makes open questions discussable without giving a legal guarantee.

Sources

  1. Muziekbedrijven Sony en Warner klagen AI-bedrijf Anthropic aan om auteursrecht ↗ · NOS
  2. Copyright and Artificial Intelligence ↗ · U.S. Copyright Office