Concerns of Microsoft and OpenAI Executives Regarding Artificial Intelligence Training Revealed
Declassified court documents submitted to the court show that tech giants knew in advance that artificial intelligence applications would harm the publishing industry.
New court documents emerging from the copyright lawsuit filed by The New York Times and other publishers have highlighted private concerns expressed by Microsoft and OpenAI executives that artificial intelligence training practices threatened the journalism industry.
Declassified Court Documents
Documents included in the summary judgment motion filed as part of the copyright lawsuit brought by The New York Times and other publishers brought the internal correspondence of technology companies to light.
The disclosed information details the potential impacts of the methods used by industry giants while training artificial intelligence models on the press and publishing sector.
Warnings from a Microsoft Executive
Brent Hecht, director of applied sciences at Microsoft, described the wholesale ingestion of news content as an unprecedented appropriation of human labor.
Hecht repeatedly warned his colleagues about the dangers of using copyrighted content to build large language models.
OpenAI Internal Correspondence and Traffic Loss
Nick Turley, who leads ChatGPT within OpenAI, stated that commercial artificial intelligence products pose a fundamental threat to the survival of publishers.
Microsoft documents discussed a destructive cycle in which declining traffic due to market substitution could deprive models of a reliable flow of information.
Paywall Obstacles and Legal Process
In the unredacted files, an OpenAI employee was seen relaying a technical solution designed to bypass publisher paywalls to OpenAI president Greg Brockman.
News organizations argue that these internal messages prove the companies knew their actions were harmful, stating that the lawsuit continues in federal court.