Skip to main content

full timeline

Major Publishers and The New York Times Intensify Legal Battle Over AI Training Data

Newly unsealed internal communications show that OpenAI and Microsoft leadership acknowledged using pirated books to train AI models and anticipated the economic disruption to human authors.

Factlen Newsroom· Updated

Follow Major Publishers and The New York Times Intensify Legal Battle Over AI Training Data in the Factlen app and get an alert when there is news. Get the app

Unsealed filings in Manhattan federal court have revealed internal communications from OpenAI and Microsoft executives.From: OpenAI and Microsoft Executives Knew Book Piracy Was Illegal, Unsealed Court Filings Reveal

Share this story

Key numbers

400
Local newspapers in new lawsuit
285,000
CPU cores in Microsoft AI supercomputer
10,000
GPUs in Microsoft AI supercomputer

From: Major Publishers and The New York Times Intensify Legal Battle Over AI Training Data

Charts, maps and images

The legal challenges highlight both the volume of publishers involved and the massive computing infrastructure required for AI training.From: Major Publishers and The New York Times Intensify Legal Battle Over AI Training Data ·
Internal messages show engineers acknowledged using pirated datasets to train early language models.From: OpenAI and Microsoft Executives Knew Book Piracy Was Illegal, Unsealed Court Filings Reveal ·
The Authors Guild argues that unchecked AI training poses an existential threat to the publishing industry.From: OpenAI and Microsoft Executives Knew Book Piracy Was Illegal, Unsealed Court Filings Reveal ·
The New York Times alleges Microsoft built a bespoke supercomputing system specifically designed to ingest copyrighted material at scale.From: Major Publishers and The New York Times Intensify Legal Battle Over AI Training Data ·