Microsoft exec called AI scraping ‘the largest theft of labor in human history,' new unredacted filings reveal
News Source
•Thu, 17 Sep 2026 19:46:08 +0000
📰 What Happened
Newly unredacted court papers in the New York Times lawsuit against OpenAI and Microsoft include a blunt admission about AI scraping. A top Microsoft executive privately called the companies’ AI training practices “theft”. OpenAI leaders reportedly said their models were an “existential threat” to the publishers and writers whose work trained them.
The filing says the companies got the material by slipping past paywalls without being noticed, building training sets through mass scraping, and stripping copyright notices out of the data. Much of the new detail comes from the Times’ own brief, and the quotes are shown without their original context.
🔍 The Backstory
The Times sued OpenAI and Microsoft three years ago. It says the firms broke copyright law by training AI models on its articles. Copyright gives writers and artists control over their work.
Courts have not settled whether training AI on copyrighted material is allowed. Judges have mostly accepted AI companies’ argument that it counts as “fair use”, which lets people use protected work in certain cases. The new material is the latest step in a case that could shape how AI learns from the web.
🎯 Why It Matters
News costs money to make. If AI companies can take it for free, fewer people get paid to report the stories you read.
Newly unredacted court papers in the New York Times lawsuit against OpenAI and Microsoft include a blunt admission about AI scraping. A top Microsoft executive privately called the companies’ AI training practices “theft”. OpenAI leaders reportedly said their models were an “existential threat” to the publishers and writers whose work trained them.
The filing says the companies got the material by slipping past paywalls without being noticed, building training sets through mass scraping, and stripping copyright notices out of the data. Much of the new detail comes from the Times’ own brief, and the quotes are shown without their original context.
The Times sued OpenAI and Microsoft three years ago. It says the firms broke copyright law by training AI models on its articles. Copyright gives writers and artists control over their work.
Courts have not settled whether training AI on copyrighted material is allowed. Judges have mostly accepted AI companies’ argument that it counts as “fair use”, which lets people use protected work in certain cases. The new material is the latest step in a case that could shape how AI learns from the web.
News costs money to make. If AI companies can take it for free, fewer people get paid to report the stories you read.