The New York Times builds an in-house AI skunkworks

The New York Times is officially setting up shop with generative AI—just weeks after suing the biggest players in the space.

Zach Seward, recently hired to head the publication’s AI initiatives, announced on Threads that the Times is building a dedicated team inside its newsroom. Their mission is to prototype uses of generative AI and machine-learning techniques to assist with reporting and reimagine how stories are presented to readers.

A skunkworks unit for journalism

According to Seward, the outlet is staffing up with a lean, cross-functional crew. The hiring plan includes a machine learning engineer, a software engineer, a designer, and a couple of editors.

Job listings are already live for an associate editorial director of AI initiatives and a senior design editor. The listing describes the group as a kind of skunkworks team within the newsroom, designed to collaborate with news, product, and tech departments to take promising prototypes into production.

At the same time, the paper is drawing a very firm line in the sand. An internal memo made clear that Times journalism will always be reported, written, and edited by its expert journalists. The AI is here to assist behind the scenes, not to replace the byline.

Suing in court, building in-house

This move comes against a backdrop of deep friction between the Times and the AI sector.

The publication was among the first major outlets to block OpenAI’s crawlers, a move that eventually snowballed into a landmark copyright lawsuit against OpenAI and Microsoft. The suit claims ChatGPT reproduces Times articles verbatim, threatening its subscription business and reader relationships.

Meanwhile, competitors have chosen a different playbook. Axel Springer and the Associated Press struck licensing deals with OpenAI to share archives and experiment with AI tools. The Times, however, is taking an adversarial approach in court while building its own experimental muscle internally. It remains unclear whether the outlet will partner with an external model vendor or develop tools purely in-house.

Why it matters

We think this is a smart, pragmatic pivot from the Times.

Fighting in court over intellectual property does not mean you should ignore the productivity shift happening under your nose. If traditional outlets refuse to touch AI out of spite, they risk falling behind digitally savvy competitors. By keeping AI strictly confined to prototyping, research assistance, and reader-facing presentation, the Times preserves journalistic credibility while still reaping technical gains.

For the media industry at large, this skunkworks experiment could set a healthier standard. Early newsroom attempts at generative AI mostly led to embarrassing errors and fake bylines. Watching one of the world’s most respected newsrooms build guarded, editor-led workflows might finally show the rest of the press how to use machine learning responsibly without gutting the newsroom.