Microsoft Exec Warns of AI-Driven 'Doom Loop' Threatening Web Performance

A recent court filing has revealed that Microsoft's director of applied science, Brent Hecht, predicted a 'doom loop' could occur due to AI models scraping and repurposing journalistic work without permission. This prediction was made in December 2023, as part of an internal presentation. The 'doom loop' refers to the situation where AI engines harm their own performance and that of the entire web by relying on scraped content.

Internal data from Microsoft suggests that Copilot's answer engine caused click-through rates for The New York Times search results to drop between 87% to 93% compared to standard Bing searches. Similarly, for Ziff Davis domains, which own Eurogamer and IGN, click-through rates dropped between 51% to 94%. Hecht described this dynamic as 'the largest theft of labor in human history.'

The court filing also reveals that when a researcher for OpenAI, Nick Ryder, suggested a way to bypass The New York Times' paywall to help scrape its writing, OpenAI president Greg Brockman replied 'ah nice.' Internal documents show Brockman writing that LLMs are 'very good at any news task.'