Microsoft staff asked if AI scraping was 'largest theft of labor'

Microsoft employees discussed whether OpenAI's use of news articles amounted to "the largest theft of labor in human history." The discussions warned that the practice could create a "doom loop" that degraded the models they were building. Court documents containing the discussions were unsealed on Thursday. The documents came from a lawsuit brought by The New York Times against Microsoft and OpenAI in late 2023. Eleven other publishers later joined the lawsuit. OpenAI has contested the claims. The case has forced OpenAI to preserve 20 million ChatGPT conversation logs. Judge Sidney Stein of the Southern District of New York is weighing summary judgment motions. A 2023 Microsoft memo said large models could be viewed as causing an unprecedented theft of labor. The memo also said large AI models destroy their supply chain. Microsoft said Brent Hecht wrote the memos. Microsoft said Hecht was a director of applied science and held a post at Northwestern University. Microsoft said Hecht was not a decision maker. Microsoft said he was employed to present divergent and asymmetric perspectives. Satya Nadella testified that anything behind a paywall should be licensed by anyone who wants to use it. Nadella said he would have exercised Microsoft's right to make OpenAI retrain its models if he had known OpenAI was training on paywalled content. A Microsoft spokesman said Nadella was speaking to broad principles about how people find and consume information. An OpenAI staffer told president Greg Brockman about building a hack to bypass The New York Times paywall. Brockman replied, "ah nice." Nick Turley wrote that AI posed an "existential threat" to publishers. Turley also wrote that AI products would become more substitutive as they improved. An OpenAI engineer wrote that users would not click links regardless of how prominently they were shown. Jack Clark warned in a 2020 memo that OpenAI was creating systems that substitute for the labor of people who define society's culture. Microsoft and OpenAI argue that the training was fair use because it transforms articles into new work rather than substituting for the originals. Steven Lieberman, who represents the New York Daily News and seven other papers, said the documents showed what the companies thought about the fairness of their conduct. The New York Times declined to comment. OpenAI did not respond to requests for comment.