NYT alleges Microsoft, OpenAI knew using news content was theft

The New York Times alleges that artificial intelligence companies OpenAI and Microsoft engaged in an unprecedented theft of copyright material by utilizing millions of news articles to train their models. According to court documents unsealed on Thursday, OpenAI scraped content from over 10 million articles, with nearly one-third sourced from the New York Times alone.

Microsoft's Director of Applied Science, Brent Hect, reportedly described this as an "astonishing theft of unprecedented proportions" and possibly the "largest theft of labor in human history." Hect is said to have warned of a potential cover-up in OpenAI's efforts to identify content originating from the New York Times and other plaintiffs.

The newspaper initiated legal action against OpenAI and Microsoft in New York federal court three years ago, alleging that the companies had stolen its copyrighted material for training OpenAI's flagship model, ChatGPT. Microsoft had invested in OpenAI in 2019. Besides the New York Times, other news publishers, such as CNET, Mother Jones, and The Intercept, have also joined the lawsuit, seeking damages for each article allegedly stolen and used by OpenAI's models.

While OpenAI has signed content licensing agreements with numerous news publishers globally since launching ChatGPT in late 2022, doubts remain about the impact of generative AI software on internet news traffic. An OpenAI engineer admitted in the court document that "no matter how prominently we show the links, users won't click."

Microsoft and OpenAI contend that their use of news content constitutes transformative work under fair use laws. In early September, the US Department of Justice filed a brief supporting OpenAI and Microsoft, citing scientific progress, economic growth, and national security as arguments for their defense. The plaintiffs have requested a summary judgment in their favor, with no trial date set until 2027.