AI Industry Execs Describe Large Language Models as 'Astonishing Theft'

A recently unredacted court filing in the New York Times v. OpenAI/Microsoft case has shed light on internal communications within the AI industry, revealing concerns about the impact of large language models on content creators.

Millions of people around the world may consider large models 'hoovering up' all their work to be an astonishing theft of unprecedented proportions, wrote Brent Hecht, Microsoft's director of applied science. This concern is echoed in internal communications within Microsoft, where they identify the 'real risk' that generative AI could significantly disrupt the employment of those who generated the data on which the foundation model was trained.

A policy document at Microsoft describes large language models as a product that destroys its supply chain. Another internal document calls this approach a 'doom loop,' where the current strategy hurts the performance of models and the entire web simultaneously.