US Department of Justice backs fair use for AI training in landmark copyright case
decoder ·
The US Department of Justice has sided with AI companies in the consolidated copyright lawsuit brought by The New York Times against OpenAI and Microsoft, arguing that training AI models on copyrighted works constitutes fair use. The DOJ distinguishes between copying works during training—which is not publicly exposed—and the model's outputs, which often lack substantial similarity to the original text. It invoked the example of Joan Didion copying Hemingway's stories to learn writing, suggesting that treating training as infringement would stifle creativity. The DOJ also contends that language models provide creative and social value, and that imposing liability would undermine the purpose of copyright law. This position directly contradicts the US Copyright Office, whose recent report rejected blanket fair use for AI training and emphasized the scale and commercial competition involved. The DOJ filing dismisses that report as non-binding and legally incorrect. The case is seen as a bellwether for AI copyright disputes.
decoder