OpenAI and Microsoft Admits LLMs Are Destroying the Web and Built on Theft
3 hours ago
- Microsoft and OpenAI executives admitted that large language models were built on an 'astonishing theft of unprecedented proportions' and the 'largest theft of labor in human history'.
- Internal documents revealed generative AI products create a 'doom loop' that kills the entire web by cannibalizing clicks from content creators.
- The admissions were part of an unredacted court filing in the New York Times vs OpenAI copyright lawsuit, showing that AI companies knew their models were trained on stolen content.
- Microsoft's CEO testified that after ripping content from news sites, clicks to those sites fell by more than 90% on Bing.
- OpenAI created a hack to bypass the New York Times paywall, and executives acknowledged that their systems substitute for human labor and destroy their own supply chain.
- The filing undermines the companies' fair use defense, as internal statements confirm they knew their products were built on theft and were destroying the internet.