Microsoft, OpenAI Workers Expressed Concerns Over AI Training Data
Internal documents reveal worries about 'theft of labor' from copyrighted content used to train AI models.
Employees at Microsoft and OpenAI have privately voiced concerns that the use of millions of news articles to develop artificial intelligence models could constitute the “largest theft of labor in human history.” These internal discussions emerged in newly unsealed court documents related to a lawsuit filed by The New York Times.
One internal Microsoft document from 2023 reportedly warned that the practice of AI models “hoovering up” copyrighted work could be seen by millions as “an astonishing theft of unprecedented proportions.” The documents were revealed as part of The New York Times’ ongoing legal case against Microsoft and OpenAI.
Separately, OpenAI's head of ChatGPT acknowledged that the company's products pose an “existential threat” to publishers, according to another internal document cited in the filings. Despite these internal reservations, Microsoft and OpenAI have publicly defended their data-gathering practices, arguing that their AI models transform scraped news content into entirely new material, thereby not violating copyright laws.