‘Doom Loop’: OpenAI and Microsoft Admits LLMs Are Destroying the Web and Built on Theft
Unsealed New York Times v. OpenAI filings show executives admitting LLMs were trained on copyrighted content and are gutting publishers' traffic.
An unredacted summary-judgment filing in the New York Times v. OpenAI copyright case exposes internal admissions from Microsoft and OpenAI. A Microsoft document called LLM training 'an astonishing theft of unprecedented proportions' and warned its AI content strategy created a 'doom loop' threatening the web's content supply chain. CEO Satya Nadella testified under oath that clicks to news sites fell by more than 90 percent on Bing after content was used for AI answers. The filing also cites an OpenAI workaround for the NYT paywall and internal descriptions of OpenAI as an 'existential threat' to news publishers.
- Unsealed summary-judgment filing in New York Times v. OpenAI quotes executives admitting LLM training used copyrighted content.
- Internal Microsoft document warns its AI content strategy started a 'doom loop' hurting models and the web.
- Satya Nadella testified Bing referrals to news sites fell by more than 90 percent after AI answers.
- OpenAI built a hack to bypass the NYT paywall; Greg Brockman responded 'ah, nice'.
- OpenAI internally called itself an 'existential threat' to news publishers.
- Victims
- The New York Times
- AI models
- GPT
From the original article
4 min read at 404media.co
"Millions of people around the world will soon consider large models ‘hoovering up’ all their work to be an astonishing theft of unprecedented proportions."
By Jason Koebler · 404media.co
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.