NYT v. OpenAI copyright arguments test AI fair use defense
Is this a scandal?
Not yet — activity is spiking. Noise 52/100, holding steady, across 1 source.
The court will likely issue a split ruling distinguishing transformative training from verbatim memorization because the Times has demonstrated specific output substitution that differs materially from the abstract training defenses accepted in Kadrey.
Noise 52/100 — louder than 99% of tracked AI controversies.
Why it matters
A ruling here will likely establish binding precedent on whether training LLMs on copyrighted news constitutes fair use or commercial infringement, fundamentally reshaping data licensing economics for the entire generative AI sector.
Key points
- NYT, OpenAI, and Microsoft filed summary judgment motions seeking resolution before a potential trial.
- The Times alleges defendants engaged in mass copying to build direct commercial substitutes for its journalism.
- OpenAI cites Kadrey v. Meta and Bartz v. Anthropic precedents claiming AI training is transformative fair use.
- Defendants argue ChatGPT outputs are not market substitutes and reported facts lack copyright protection.
- Microsoft advanced identical fair use defenses regarding its Copilot AI product's training methodology.
- The Times does not dispute the specific fair use findings in the cited California cases but distinguishes its claims.
The story
The New York Times, OpenAI, and Microsoft presented summary judgment arguments Friday in a federal copyright lawsuit that could determine the legality of training large language models on protected journalism. The Times alleges the defendants copied its works at scale to create commercial substitutes, violating copyright law. OpenAI and Microsoft counter that training on publicly available content is transformative fair use, citing recent California rulings in Kadrey v. Meta and Bartz v. Anthropic that favored AI developers. The defendants further argue that ChatGPT outputs do not function as market replacements for Times articles and that reported facts remain unprotected. A judicial decision favoring the Times could force AI companies to overhaul data sourcing strategies, while a victory for OpenAI would significantly weaken publishers' ability to monetize content used for model training. Both parties seek a favorable ruling before a potential trial.
Who's involved
Argues OpenAI and Microsoft copied works at scale to create illegal commercial substitutes that violate copyright law.
Contends that training on public works is transformative fair use and that outputs do not serve as market substitutes.
Asserts Copilot's training methodology mirrors OpenAI's protected fair use practices and respects factual reporting exemptions.
Noise Level
The timeline
Parties present summary judgment arguments
NYT, OpenAI, and Microsoft filed motions asking the judge to rule on core liability issues ahead of potential trial.
- 2025-00-00T00:00:00Z
Bartz v. Anthropic affirms fair use
Court ruled Anthropic's use of copyrighted texts for Claude training was transformative and non-infringing.
- 2025-00-00T00:00:00Z
Bartz v. Anthropic ruling issued
Court found Anthropic's training on copyrighted texts was transformative, reinforcing the fair use defense.
- 2024-08-00T00:00:00Z
Kadrey v. Meta ruling favors AI training
California federal court found using copyrighted books to train Llama models constituted transformative fair use.
- 2024-08-00T00:00:00Z
Kadrey v. Meta ruling issued
California federal court ruled using copyrighted books to train Llama models constituted transformative fair use.
NYT files landmark copyright lawsuit
The New York Times sued OpenAI and Microsoft in federal court alleging unauthorized use of millions of articles for AI training.
NYT files copyright lawsuit
The New York Times sued OpenAI and Microsoft alleging unauthorized use of millions of articles for AI training.
- 2 days ago
Summary judgment arguments presented
All three parties submitted official motions arguing for case resolution ahead of a potential jury trial.
The full record
Sources & methodology
- Historic NYT v. OpenAI copyright battle heats up — axios.com
Every claim above traces to these primary items. How we score →
The forecast
The court will likely issue a split ruling distinguishing transformative training from verbatim memorization because the Times has demonstrated specific output substitution that differs materially from the abstract training defenses accepted in Kadrey.
Forecast, not fact — an editorial estimate we score when this resolves.
That's the complete picture as of — nothing more to know right now. We'll update this page the moment it changes.
Follow this story
We keep this page current — no need to check back. We'll send the next real change to your inbox, nothing else.
Tracking this story since September 8, 2026.
Join the Discussion
Discuss this story
Community comments coming in a future update
Be the first to share your perspective. Subscribe to comment.