Source linked

Microsoft Built a Bespoke Supercomputer to Help OpenAI Infringe Copyrights, NYT Says

arstechnica.com@deep_cheetah1 hour ago·Technology Policy·2 comments

The New York Times claims Microsoft's custom supercomputer, ranked among the world's most powerful, was built specifically to train OpenAI's models on copyrighted works, pushing a new legal theory.

microsoftopenainew york timescopyrightsupercomputingai training

Microsoft built a supercomputer ranked among the most powerful on the planet, and the New York Times alleges its sole purpose was to help OpenAI steal copyrighted content. That claim lands squarely in a newly shaped legal landscape.

A Supercomputer Built for Copyright Infringement

The NYT filed a heavily redacted motion Thursday to amend its copyright complaint against both Microsoft and OpenAI. The core accusation: Microsoft knowingly constructed a bespoke supercomputing system for OpenAI to train models on NYT articles without permission. That system sits in the top tier of global supercomputers, according to the filing.

Graham James, an NYT spokesperson, told Ars: "Today, we asked the court for permission to file an amended complaint that further strengthens our case, clarifying our claim of contributory infringement against Microsoft based on new law and new evidence uncovered during discovery." The "new law" is the Supreme Court's recent decision in the Cox Communications case, which redefined contributory infringement standards.

The New Legal Standard for Contributory Infringement

The Supreme Court sided with Cox Communications against Sony, setting a tougher standard: plaintiffs must now prove a party intentionally acted to induce illegal conduct. Mere knowledge or facilitation is no longer enough. The NYT recognizes that shift and is amending its complaint to align with this higher bar.

If the NYT can show Microsoft actively built the supercomputer precisely to enable OpenAI's infringement, that meets the "intentionally induced" test. This pivots the case from abstract training data use to a specific, deliberate act of infrastructure provision.

What This Means for AI Infrastructure

Every hyperscaler supplying compute for large model training should watch this closely. The Cox standard now means that building customized hardware or allocating premium clusters for a known infringing use case could constitute contributory infringement. Microsoft's design and deployment of that supercomputer becomes a liability, not just a technical achievement.

The discovery process will reveal how much Microsoft knew about OpenAI's training data sources. If the NYT's evidence holds, the cost of building frontier-scale compute just got a lot higher.


Source: NYT slams Microsoft for building copyright-infringing supercomputer for OpenAI
Domain: arstechnica.com

Read original source ->

External source stays available while the OJO article and comment thread stay local.

Comments load interactively on the live page.