The New York Times has filed an amended complaint in its copyright infringement lawsuit against OpenAI and Microsoft, first lodged in December 2023 in federal court in New York. The newspaper alleges the companies used millions of Times articles without permission to train large language models, which now compete directly with the Times as an information source.
OpenAI and Microsoft have integrated those models into products like Copilot, deriving commercial benefit. The Times presented evidence that GPT-4 outputs reproduced content nearly verbatim and even fabricated quotes attributed to real articles. A judge has allowed core copyright claims to proceed, dismissing some Digital Millennium Copyright Act counts.
OpenAI argues fair use, claiming training on publicly accessible text is transformative. The Times counters that near-exact reproduction is not. The outcome holds implications for every major AI developer: a Times victory opens the door for other publishers to seek compensation; a ruling for OpenAI would ratify the current training model.
With billions at stake, the case is the most factually developed challenge to AI training practices.