Dwi Wanna
In the high-stakes theater of modern intellectual property law, few battles carry as much weight as the ongoing litigation between The New York Times (NYT) and the tech giants OpenAI and Microsoft. Initiated in late 2023, the lawsuit—which alleges massive copyright infringement through the training of Large Language Models (LLMs)—has transcended a mere business dispute to become a definitive test case for the future of journalism, creativity, and artificial intelligence.
As the legal proceedings stretch into their third year, a new chapter has emerged. The Times has recently petitioned the court for a summary judgment, armed with a revealing legal brief that draws upon internal documents and statements from the defendants. While much of this evidence remains shielded behind court-ordered seals and redactions, reporting from outlets like 404 Media suggests that the internal discourse within OpenAI and Microsoft reveals a profound, if not cynical, understanding of the existential threat their own technologies pose to the digital ecosystem.
Chronology of the Conflict: From Collaboration to Courtroom
The relationship between media conglomerates and AI developers began with a veneer of curiosity, but quickly curdled into open hostility.
- Pre-2023: For years, OpenAI and Microsoft scraped vast swaths of the public internet—including millions of articles from The New York Times—to feed the hungry algorithms of GPT-3, GPT-4, and the Copilot suite. At the time, this was often framed as "fair use" for the purpose of innovation.
- December 2023: The New York Times formally filed its complaint in the Southern District of New York. The lawsuit alleged that the defendants’ models were built by copying millions of the Times’ articles without permission or compensation, and that these models now compete directly with the publication by providing users with verbatim or near-verbatim summaries of its journalism.
- 2024: The discovery phase of the trial began, characterized by intense legal maneuvering. Both Microsoft and OpenAI sought to protect their trade secrets, successfully petitioning the court to keep significant portions of their internal communication logs and training data methodology under seal.
- 2025–2026 (The Current Status): The legal battle has moved into a more aggressive phase. The Times’ recent move for a summary judgment indicates that they believe the evidence gathered during discovery is sufficiently damning to win the case without a full trial. They are effectively asking the court to rule that, as a matter of law, the defendants’ actions constitute infringement that cannot be excused by "fair use" arguments.
The "Doom Loop": Internal Admissions and the Theft of Labor
Perhaps the most explosive element of the recent filing is the contention that OpenAI and Microsoft’s own leadership recognized the parasitic nature of their business model. According to reports, internal communications reveal a stark acknowledgment that the very web content required to build functional LLMs is being systematically "destroyed" by those same models.
This phenomenon, often dubbed the "Doom Loop," posits that AI models rely on high-quality human journalism to remain accurate and relevant. However, by surfacing that information directly to users, AI tools reduce the incentive for users to visit the source websites. As traffic drops, publishers lose the advertising revenue necessary to pay the journalists who create the content in the first place. The AI eventually runs out of new, high-quality information to ingest, leading to a decline in model performance.
The legal brief submitted by the Times reportedly highlights internal rhetoric from OpenAI and Microsoft executives describing the scraping of copyrighted labor as the "biggest theft in human history." These admissions, if substantiated in open court, could dismantle the defendants’ primary defense: that their data ingestion is transformative and beneficial to the public good.
Supporting Data: The Mechanics of Infringement
The Times’ argument is not merely philosophical; it is grounded in technical evidence of how ChatGPT and Copilot interact with paywalled content.
The "Verbatim" Problem
One of the core grievances in the complaint is the tendency for models to regurgitate chunks of articles that are otherwise locked behind a paywall. By providing these snippets, the AI removes the user’s need to subscribe to the Times, directly cannibalizing the publication’s subscription-based business model.
The Attribution Gap
The lawsuit points out that when AI models summarize news, they often strip away the branding, advertising, and context that the Times relies on for its brand identity. This creates a "hallucination" of authority, where the AI presents the information as its own, thereby diluting the value of the original journalistic entity.
The Economic Scale
While the exact dollar amount of damages has not been finalized, the Times is seeking potentially billions in compensation. The sheer scale of the infringement—involving millions of articles scraped over a multi-year period—suggests that the outcome of this case will set a precedent for how every other media outlet in the world negotiates with AI companies.
Official Responses: The Battle of Legal Philosophies
The defendants have maintained a consistent, albeit defensive, stance throughout the litigation.
OpenAI’s Position:
OpenAI has consistently argued that its technology does not "copy" content in the traditional sense. Instead, it "learns" from the information much like a human student reads a textbook to gain knowledge. They maintain that the creation of generative AI is protected under the doctrine of "fair use," as it provides a transformative service that adds significant value to society by making information more accessible and actionable.
Microsoft’s Position:
Microsoft, which acts as both an investor in OpenAI and a primary integrator of the technology into its Windows and Office ecosystems, has attempted to distance itself from the direct act of scraping. They argue that as a platform provider, they are not responsible for the specific datasets used to train the third-party models they integrate. However, the Times has pushed back, arguing that Microsoft’s role in the development of Copilot and the deployment of OpenAI’s models makes them equally liable for the underlying copyright violations.
The New York Times’ Stance:
The Times asserts that "fair use" does not grant tech companies the right to monetize the life’s work of journalists at the expense of the industry’s survival. Their legal team has made it clear that they are not inherently anti-AI; rather, they are anti-theft. They seek a licensing framework where AI developers must pay to use professional, vetted content—a move that would establish a "copyright tax" on the future of AI.
The Broader Implications: A Crossroads for the Internet
The outcome of this lawsuit will likely dictate the structure of the internet for the next two decades.
1. The Death of the "Open" Web?
If the courts rule in favor of the Times, we may see a "walled garden" future. Publishers may implement technical barriers (such as advanced bot-detection or encrypted content) that prevent AI crawlers from accessing their sites. This could effectively end the era of the "open" web, as information becomes increasingly siloed behind AI-exclusive licensing agreements.
2. The Economic Survival of Journalism
If the defendants prevail, it could signal the end of traditional news organizations as viable business entities. If tech giants can legally train models on copyrighted works without compensation, the incentive to produce high-quality, long-form investigative journalism diminishes. The result would be an ecosystem dominated by AI-generated "slop," lacking the verification, ethics, and depth of human-led reporting.
3. Regulatory Precedent
This case is currently the bellwether for AI regulation. Legislators in the EU, UK, and beyond are watching the New York court with bated breath. A victory for the Times would provide the legal roadmap for other industries—such as music, film, and art—to demand compensation for their own datasets.
Conclusion: A Fight for the Future of Truth
The summary judgment request by The New York Times represents a pivotal moment in the history of the information age. By forcing the hand of OpenAI and Microsoft, the Times is compelling the tech industry to justify its growth in the face of mounting evidence that it is built on a foundation of intellectual property exploitation.
As the legal briefs continue to move through the courts, the world remains divided. To some, OpenAI and Microsoft are the pioneers of a new, post-scarcity information age. To others, they are simply the latest in a long line of corporate entities that believe they are entitled to seize the labor of others for profit. Regardless of one’s stance, the reality is clear: the era of "move fast and break things" is colliding with the reality of property rights. The verdict in this case will not just decide the fate of a newspaper; it will determine whether the internet of the future will be a vibrant ecosystem of human creativity or a sterile output of automated theft.
Jia Lissa
Lina Hope
