Internal documents from Microsoft and OpenAI, revealed in a lawsuit by The New York Times, highlight internal concerns regarding AI data scraping. Brent Hecht, a Microsoft director, described the practice as the largest theft of labor in history in a January 2023 memo. He also warned that AI models threaten the economic foundations of their essential content suppliers.
Data from Microsoft indicated that the Copilot tool reduced click-through rates for The New York Times by up to 93% compared to Bing search. Microsoft internal communications referred to this phenomenon as a doom loop that negatively impacts both model performance and the broader web ecosystem. The company acknowledged that creators are neither compensated nor intended for their content to be used this way.
OpenAI leadership also expressed concerns, with the head of ChatGPT, Nick Turley, labeling the chatbot an existential threat to publishers due to its substitutive nature. Internal messages showed OpenAI staff discussing methods to bypass the paywall of The New York Times. Meanwhile, CEO Satya Nadella stated that paywalled content should be licensed for training purposes.
AI companies maintain that their data scraping practices constitute fair use under legal standards. However, the newly disclosed internal communications could challenge this defense. The legal proceedings focus on whether the commercial nature of these models and their impact on the market value of copyrighted works align with existing fair use criteria.
