NYT amends suit: Microsoft built supercomputer to help OpenAI steal articles
New evidence alleges Microsoft custom-built a supercomputer to train AI on NYT content without permission.
Deep Dive
The New York Times moved to amend its copyright suit against Microsoft and OpenAI, claiming Microsoft built a bespoke supercomputer to train AI on copyrighted articles. The amendment follows a Supreme Court ruling on contributory infringement standards. NYT alleges Microsoft's system disproportionately weighted its articles, and that deploying Times-trained LLMs helped boost Microsoft’s market capitalization by a trillion dollars in the past year. NYT also dismissed two claims to streamline the case.
Key Points
- NYT claims Microsoft built a custom supercomputer to train OpenAI's models on copyrighted NYT articles, not just generic cloud services.
- The amendment responds to a Supreme Court ruling that tightened contributory infringement standards, requiring proof of intentional inducement.
- NYT dropped two claims (trademark dilution and one contributory infringement count) to streamline its case while adding specificity about Microsoft's role.
Why It Matters
This case could set a precedent for AI training on copyrighted data and big tech's liability for custom infrastructure.