Times Publisher Defends Newsroom Against AI Threats and Government Pressure

The New York Times is spending north of $20 million to sue OpenAI and Microsoft over copyright infringement—a costly gamble the company's publisher, A.G. Sulzberger, frames as a necessary defense of journalism itself. Speaking in a recent interview, Sulzberger outlined a dual crisis threatening the news industry: an administration wielding legal intimidation against reporters and tech companies freely extracting journalistic content to train AI systems without compensation.

The Administration's Attack on Press Freedom#

Sulzberger characterized recent subpoenas issued by the Trump Justice Department to Times journalists covering a Qatar-donated jet as part of what he called the most aggressive assault on the press from the White House in over a century. The department went beyond targeting the reporters themselves, seeking phone records and text messages from their family members—a move Sulzberger called "a gross overreach."

He emphasized that other news organizations have largely maintained their commitment to publishing truth despite pressure, citing examples like the Associated Press and Wall Street Journal defending their independence. However, he acknowledged concern about a subtler threat: reporters at competing outlets privately expressing reluctance to break stories for fear that editors will suppress coverage out of retaliation anxiety. This chilling effect, he argued, represents the real danger to a free press.

AI as a Reporting Tool—With Guardrails#

Rather than embracing generative AI for writing, Sulzberger described the Times' approach as experimental and cautious. The newsroom uses AI primarily to identify patterns and anomalies in large datasets—work fundamentally suited to computational speed. One Pulitzer Prize-winning reporter told Sulzberger that AI works best as a "filter," capable of narrowing millions of images down to thousands, but requiring the journalist to verify every result.

Sulzberger was explicit about responsibility: humans remain accountable for everything published. Some news organizations made "incredibly foolish" decisions, he suggested, outsourcing writing to AI and inventing fictional bylines—a cascade of preventable errors that underscored AI's status as a tool, not a replacement for editorial judgment.

The 'Original Sin' of AI Training#

Yet Sulzberger's gravest concern centers on how AI companies built their products. While coders command "sports-star level" compensation and companies spend hundreds of billions on infrastructure, he said the industry simply "took" journalistic content without permission or payment. This asymmetry—where the most valuable companies in history capture and privatize the work of struggling news organizations—threatens journalism's economic survival.