11 hours ago
5
High-level staffers at a pair of top tech companies knew that scraping online news content to bolster their AI models could have devastating impacts on publishers – but that didn’t stop them, according to a bombshell legal filing.
One OpenAI executive even warned that the shady tactic, in which tech giants circumvent paywalls to access stories to train their AI, could pose an “existential threat” to news organizations, the Wall Street Journal reported, citing the newly unsealed court documents in a massive copyright case.
A director at Microsoft had also sounded the alarm over the dodgy practice.
“Our AI content strategy has started a ‘doom loop’ that will hurt the performance of our models and the entire web at the same time,” said Brent Hecht, Microsoft’s director of applied science, the filing shows.
“It is highly unusual that an end-product threatens the economic foundations of its essential suppliers, but that is the situation we have created for our LLM business with respect to its ‘content supply chain,’” said Hecht.
The filing was made public Thursday in a copyright infringement case brought against Microsoft and OpenAI by several news publishers, including The New York Times.
The Times sued OpenAI and Microsoft in December 2023, alleging that ChatGPT and Microsoft’s Copilot used millions of their articles to train its models and answer users’ questions.
Other papers owned by Alden Global Capital, such as the Chicago Tribune and the New York Daily News, also sued the companies in 2024, but the cases were eventually consolidated, the outlet reported.
ChatGPT chief Nick Turley acknowledged that AI products could become increasingly “substitutive” and said that they posed an “existential threat” to publishers, the filing alleged.
AI products “will get more and more substitutive as they get better,” Turley wrote, per the outlet.
OpenAI internal documents even referred to ChatGPT as a “modern newsstand,” the filing said.
The document also detailed internal discussions about skirting publishers’ paywalls — including an exchange in which OpenAI President Greg Brockman responded “ah nice” after a researcher found a hack to get around the Times paywall.
The news organizations are looking for a federal judge to rule in their favor before the case goes to trial, citing internal documents and testimony from Microsoft CEO Satya Nadella, Brockman, and Turley.
The publishers detailed the allegations against the top AI companies in the brief that was jointly filed.
Nadella had testified that “anything that is paywalled should be licensed by anyone who wants to use it,” for AI training, adding that he would have Microsoft force OpenAI to retrain its models after learning they had used paywalled content.
The Microsoft CEO also testified that conversing with the chatbots could be a substitute for websites, instead of “needing to go to the underlying source,” the document detailed.
The tech companies have contended scraping from news sites constitutes “fair use,” allowing some copyright material to be used without explicit permission.
Hecht’s comments “reflect one employee’s individual perspective, are not a legal analysis, and do not represent the company’s views,” a Microsoft representative told The Post.
“Microsoft’s position is set out in its court filings, which explain why these transformative uses are consistent with copyright law and why Copilot is not a substitute for publishers’ journalism,” the spokesperson continued.
The Justice Department also recently filed a statement of interest in the suit, arguing that the companies’ use of news content is allowed under fair use in US copyright law, the outlet reported.
OpenAI did not immediately respond to The Post’s request for comment.
English (US)