Who owns what a model learned?
Every modern generative model is trained on vast collections of text, journalism and literature. The legal fight isn't really about the model's answers — it's about how the model was trained and who owns the knowledge it absorbed.
One distinction runs through every case, and it's the spine of this casebook: courts increasingly treat the training of a model differently from how the data was acquired. Learning from a work can be lawful even when the way the work was obtained is not.
Lawful acquisition + training on one side. Piracy on the other. Almost every hard case is an argument about which side a fact pattern lands on.
What you'll be able to do
- 1Separate acquisition, training and output when spotting the copyright question.
- 2Use fair use and market substitution to explain the competing arguments.
- 3Explain Bartz v. Anthropic without turning it into a universal “training is legal” rule.
- 4Tell allegations and advocacy apart from rulings in the NYT litigation.
- 5See why dataset provenance is becoming the template for all scraped data.
Creators vs. developers
Creators say…
Feeding our work into a model at scale is unauthorized exploitation. No one asked — and the model competes with us using what it took.
Developers say…
Training is transformative learning. The model studies patterns; it doesn't republish the book. That's classic fair use.
Two ideas decide most of these fights — you'll weigh them yourself on the next page:
The fair-use balance
Toggle each factor on or off and watch the scale tip. There's no single switch — courts weigh everything together. Find a mix that leans fair use, then one that leans infringement.
Which factor most weakens a fair-use defense?
Three phases of the dispute
The conflict moved through three stages. Reconstruct the record — click them from earliest to latest.
Click the earliest phase first.
Bartz v. Anthropic
The landmark outcome so far — and it drew the key line in plain view. Judge Alsup split the question in two:
That unlawful acquisition created serious liability and drove a historic $1.5 billion settlement — the largest copyright class settlement in U.S. history, finalized in July.
Where does liability enter?
Follow one dataset through the pipeline. Click each stage — the question at every node is the same: was this step lawful?
In Bartz v. Anthropic, what did Judge Alsup rule was not protected?
The New York Times v. OpenAI & Microsoft
The most-watched case still live. The publishers say their investigative journalism was ingested without permission; the tech firms say it's fair use.
Cross-motions for summary judgment were set for October 21. The DOJ filed a letter backing the fair-use defense, which the publishers are fighting. Courts are weighing technical evidence, market substitution, and fair-use defenses.
The provenance template
Bartz's split — lawful acquisition + training = fair use, but piracy = infringement — is becoming the template for how courts treat scraped data. And it reaches every kind of generator: text, images, audio, code.
That's why dataset provenance review — proving where training data came from and how it was obtained — has become a practical compliance layer. With many lawsuits active and interpretations competing, copyright law here is still contested and evolving.
The decisive question is almost always: lawful acquisition + training — or piracy?
Knowledge-check score
Fair use — a U.S. defense allowing some unlicensed uses, especially transformative ones. Transformative use — adds new purpose/character. Shadow libraries — pirate repositories of copyrighted works. Market substitution — whether an output competes with/replaces the original. Dataset provenance — the record of where training data came from and how it was obtained.