asked a judge on Thursday for permission to amend its copyright complaint against OpenAI and Microsoft, adding a sharper allegation that Microsoft built a bespoke supercomputing system to help OpenAI steal the Times’s works. The proposed filing also trims the case, with the newspaper agreeing to drop two claims against all defendants.
That move matters now because the Times says the law changed under its feet. After a Supreme Court ruling in a separate case involving Sony and Cox Communications, plaintiffs pursuing contributory infringement must show that a defendant intentionally induced illegal conduct. The Times says its revised complaint is built to meet that standard and to keep its claims focused on what it sees as the core wrongdoing: the use of its journalism in AI training and outputs without permission.
Graham James said the newspaper uncovered new evidence during discovery and was asking the court to allow the amended pleading. In a statement, he said the Times was “clarifying our claim of contributory infringement against Microsoft based on new law and new evidence uncovered during discovery,” and added that Microsoft had “actively encouraged OpenAI to steal our copyrighted works.” He said the newspaper’s core claims had not changed: Microsoft and OpenAI, he said, stole millions of the Times’s copyrighted works to compete with its products and enrich themselves.
The filing, which was heavily redacted, tries to make that accusation more specific. The updated complaint says the supercomputer was tailor-made to help OpenAI infringe and was built for the explicit purpose of training AI on copyrighted works without permission. It also says the Times’s articles were more heavily weighted by the system because both firms wanted to train models on the highest-quality journalism possible. That is the kind of detail that can matter in court, because it moves the case from a broad complaint about AI scraping to a claim about intentional design and alleged instruction.
Microsoft does not see it that way. A spokesperson said the company viewed the amended complaint as a last-ditch effort to save the Times’s claim from unfavorable precedent in other recent rulings. That split goes to the heart of the case: the Times says it has new evidence and a stronger legal theory after the Supreme Court decision, while Microsoft says the newspaper is reworking its lawsuit to avoid a bad result elsewhere.
The case has been building since 2023, when became the first major publisher to sue OpenAI. The original suit said ChatGPT was illegally trained on Times articles, reproduced them verbatim, and caused market and reputational harms. It also said summaries of Wirecutter reviews could reduce writer commissions by taking away clicks on affiliate links. By seeking to amend the complaint now, the Times is trying to keep the case aligned with the newer legal standard and narrow it to its strongest claims.
The judge now has to decide whether the new version can be filed. If it is allowed, the case will move forward with a more direct allegation that Microsoft helped build the machinery behind the infringement claim, and with two claims gone from the case entirely. If it is not, the Times will have to keep fighting with the complaint it already has.
