What is copyright infringement?
Copyright infringement is when someone uses the whole or a substantial part of your work without your permission as the copyright owner. Copyright laws ensure that your intellectual property remains yours, protecting your exclusive rights to reproduce and distribute your work. But what happens when AI challenges these rules?
Current Case: The New York Times v OpenAI and Microsoft
Last Christmas, the New York Times (NYT) filed a copyright infringement lawsuit in the Federal District Court in Manhattan against OpenAI and Microsoft. The NYT alleged that OpenAI used millions of copyrighted articles to train their large language models (LLMs). By “free-riding on the NYT’s massive investment into journalism”, OpenAI has been accused of bypassing paywalls and offering AI-generated responses that compete with NYT’s original articles.
At the heart of the case lies a critical question: Is this fair use?
OpenAI is unlikely to deny the literal copying of the NYT’s articles to train their LLMs. However, under the US Copyright Act, fair use is a statutory defence that allows certain uses of copyrighted material such as for education, commentary, or satire.
One test for fair use is whether OpenAI’s work is transformative or simply duplicative. OpenAI will likely argue that training LLMs on large datasets is transformative, as their AI doesn’t reproduce articles but creates new outputs. They could argue that reconstructions of works are not infringement if they have the same ideas but are different in their expression and not copied verbatim.
However, the most important test for fair use is the economic impact on the original work due to infringement. If OpenAI undermines the NYT’s subscriptions, the case for damages — potentially “billions of dollars” — becomes stronger.
Trial?
This issue is contentious since generative AI relies on learning from vast volumes of text —copyrighted or not. Courts now face the challenge of determining whether training an AI model inherently constitutes unlawful copying.
Whether this case will go to trial or settle quietly is still uncertain. Settling is attractive for its cost-efficiency and privacy, but a clear ruling — potentially by the Supreme Court —could set precedence and shape the future of AI and the interpretation of copyright law.
Looking Ahead
As technologies improve, more novel cases will arise, challenging existing statutory definitions. If legislation does not adapt, we are reliant on the courts to set precedence in the intellectual property sphere to accommodate technological advancement. This would rely on companies choosing to litigate even if it is the more costly option.
Bibliography:
Reuters: "Tech companies face tough AI copyright questions in 2025" Reuters
Harvard Law Review: blog post "NYT v. OpenAI: The Times's About-Face" Harvard Law Review
Kluwer Copyright Blog: The article "Is Generative AI Fair Use of Copyright Works? NYT v. OpenAI" Copyright Blog
The Washington Post: "AI's future could hinge on one thorny legal question" The Washington Post
The Guardian: "Microsoft asks to dismiss New York Times's 'doomsday' copyright lawsuit" The Guardian
Comments