The New York Times is escalating its landmark legal battle against the biggest players in artificial intelligence, shifting its crosshairs directly onto the hardware that made ChatGPT possible. In an updated complaint, the publication alleges that the Microsoft supercomputer used by OpenAI was not a generic cloud service, but a custom-built machine designed explicitly to ingest copyrighted works without permission. This aggressive pivot attempts to hold the infrastructure provider equally liable for the massive data scraping required to train modern AI.
When the NYT became the first major publisher to sue OpenAI in 2023, the initial allegations focused heavily on the outputs. The newspaper argued that ChatGPT illegally reproduced its articles verbatim, hallucinated false claims attributed to NYT reporters, and cannibalized affiliate revenue by summarizing Wirecutter reviews. However, the revised lawsuit now targets the foundational architecture, claiming Microsoft actively facilitated the infringement by building an "unusually complex" system tailored to seize copyrighted material.
Microsoft specifically designed it for the purpose of using essentially the whole Internet - curated to disproportionately feature Times Works - to train the most capable LLM in history.
- The New York Times
According to the filing, both tech giants deliberately weighted NYT articles heavily during the training process to ensure their models could confidently mimic high-quality journalism. The publication argues this strategy has been immensely profitable, noting that "Microsoft’s deployment of Times-trained LLMs throughout its product line helped boost its market capitalization by a trillion dollars in the past year alone," according to the updated complaint. To prove market substitution, the NYT presented discovery evidence showing users actively using ChatGPT to bypass paywalls by simply requesting the "next paragraph" of an article.
The End of Cloud Infrastructure Immunity
This strategic shift by the NYT represents a dangerous new frontier for tech giants providing the computing backbone for generative AI. By arguing that Microsoft’s hardware was custom-engineered specifically for copyright infringement, the lawsuit attempts to strip away the traditional "safe harbor" protections that cloud providers usually enjoy. If the court agrees that building a specialized supercomputer for a partner constitutes active participation in data theft, it could fundamentally alter how companies like Amazon, Google, and Microsoft structure their AI computing partnerships.
Furthermore, the discovery evidence showing users explicitly prompting ChatGPT to bypass paywalls provides a concrete link between AI capabilities and direct financial harm. By proving that users are actively utilizing the model as a direct substitute for a paid subscription, the NYT makes the "fair use" defense significantly harder for OpenAI to maintain in court.
