The New York Times has accused OpenAI of carrying out “an astonishing theft of unprecedented proportions” by using millions of news articles to train its artificial intelligence models, according to a court document unsealed yesterday.
The San Francisco-based AI company allegedly scraped more than 10 million articles, with almost one in three reportedly coming from The New York Times.
- Advertisement -
According to the filing, Microsoft Director of Applied Science Brent Hect described the alleged conduct as “an astonishing theft of unprecedented proportions” and potentially “the largest theft of labour in human history”.
Mr Hect also allegedly cautioned that OpenAI may have carried out an “accidental cover up” while trying to determine which material in its systems originated with The New York Times and the lawsuit’s other plaintiffs.
The newspaper sued OpenAI and Microsoft in New York federal court three years ago, accusing the companies of using its copyrighted work to train OpenAI’s flagship ChatGPT model without permission.
Microsoft made its first investment in OpenAI in 2019.
US Department of Justice backs OpenAI and Microsoft
Several other publishers later joined the case, including Ziff Davis, the owner of CNET and other technology publications, along with the parent company of Mother Jones, investigative outlet The Intercept and numerous local newspapers across the United States.
The plaintiffs are seeking damages for every article they say was taken and used by OpenAI’s models, although the potential total remains unclear.
Since ChatGPT launched in late 2022, OpenAI has reached content-licensing agreements with numerous news organisations worldwide. Even so, publishers remain concerned that generative AI tools could divert readers from news websites by providing information directly.
AI companies have attempted to address those fears by attaching citations and links to responses, but publishers and others have criticised the approach as inadequate.
“No matter how prominently we show the links, users won’t click,” one of OpenAI’s engineers reportedly acknowledged, according to the court document.
Microsoft and OpenAI maintain that their use of news material is transformative and protected under “fair use” laws.
In early September, the US Department of Justice submitted a brief supporting the companies, citing “scientific progress”, economic growth and “national security” among its arguments.
The plaintiffs have asked the court to issue summary judgment in their favour.
If US District Judge Sidney Stein grants the request, the case would be resolved without a trial. A ruling, however, is not expected until 2027.
