When the Newsroom Becomes a Server Farm

The new fake news threatening real journalism

In May 2026, public radio stations WLRN and WUWF traced the rise and fall of what appeared to be a local news outlet in South Florida. It had a logo, bylines, and a beat. It had no reporters. The articles were written by language models, the bylines were invented, and the operation existed to harvest search traffic until it was caught. For months, the town it claimed to cover was unable to distinguish whether it still had a newspaper at all.

This is just one example of the captured-and-replaced American press. It is no longer an anomaly.

For the press to fulfill its role as Edmund Burke’s “fourth estate” — an unelected, adversarial force charged with holding power to account — has always been an expensive proposition. Woodward and Bernstein spent two years on Watergate before their reporting brought down a president. Four decades later, the Guardian scattered Edward Snowden’s NSA archive across jurisdictions so no single government could kill the story.

The Watergate and Snowden exposés worked in part because producing a convincing counter-narrative was nearly as expensive as uncovering the truth. To kill a real story, opponents needed real PR machinery, real surrogates, real fake documents, and real time to coordinate. Truth was hard to produce, but so were plausible lies. That symmetry gave verified reporting a fighting chance.

AI breaks the economics of investigative journalism. The costs of reporting and verifying truth have not changed, but now chatbots collapse the cost of plausible counter-truth to near zero. The fight is suddenly far less fair.

Three forces are tipping the scales, the first of which is mechanical. Pew Research Center found that Google’s AI Overviews cut publisher click-throughs nearly in half: when an Overview appears at the top of a search results page, only 8 percent of users click a ranked link below, compared with 15 percent on pages without . Chartbeat data cited by Press Gazette indicates that global publisher traffic from Google dropped by roughly one-third in the year to November 2025.

With diminishing search revenue, publisher layoffs have risen. Business Insider cut 21 percent of its workforce in May 2025, citing AI and “extreme traffic drops” outside the company’s control. Penske Media sued Google in September 2025, alleging the company is using its search monopoly to force publishers to allow their content into AI training and Overviews; Google moved to dismiss in January 2026.

The second force is the bailout. The same AI labs whose products are siphoning traffic from publishers are paying a handful of large survivors to accept compensation rather than challenge the use of their content. News Corp’s licensing deal with OpenAI runs up to $250 million over five years, the largest single AI content agreement in publishing history. Axel Springer, owner of Business Insider, Politico, and Bild, signed for roughly $13 million a year over three years. The Atlantic, Vox, the Associated Press, the Financial Times, Le Monde, and People have signed comparable terms, most of them undisclosed.

By paying select media companies, AI firms are creating a class of newsrooms financially dependent on the very industry they are expected to cover. The deals laundered in press releases as “partnerships” are functionally subscriptions, yet at the same time it becomes dangerous territory for those outlets to critique their AI benefactors. Publisher CEOs praise the new influx as “transformative revenue” but the conflict of interest is clear. The reporters on the AI beat have not been asked to comment.

The third force is the substitute. NewsGuard, which tracks unreliable news domains, counted 1,265 “pink slime” outlets in mid-2024. These are partisan operations disguised as local papers, now increasingly written by large language models. The number is now over 1,600. With only 1,213 daily newspapers still operating in the United States, more ostensibly local outlets online are fake than real. NewsGuard’s separate AI Content Farm tracker logged 3,006 unreliable AI-generated news sites across 16 languages by March 2026, up from 49 in 2023.

The stories produced by these façades are designed to be indistinguishable from real reporting at near-zero marginal cost. Politically, they work. One pink-slime network has boasted in its own case studies that a $49,000 ad campaign across twelve Iowa counties produced roughly 3,300 additional votes. The most valuable soil into which they are planted is the news desert. Medill’s State of Local News 2025 report counts 3,500 newspapers gone in two decades, 213 counties with no local outlet, another 1,524 with only one, and 50 million Americans with limited or no access to local journalism. Where the paper used to be, a synthetic outlet has now sprung up.

Where is the government in all this? Oversight normally fall to the FTC under Andrew Ferguson, the FCC under Brendan Carr, and the DOJ’s Antitrust Division. The first two have shown little interest in the issue, and the third has been in flux since the Trump administration forced out antitrust chief Gail Slater in February 2026. None has acted on the deals between AI labs and the press. In Congress, neither Chuck Grassley’s Senate Judiciary Committtee nor Brett Guthrie’s House Energy and Commerce Committee appears to have scheduled a hearing on the issue. As of now, nobody is clearly in charge — and that, itself, is noteworthy.

The case everyone is watching is The New York Times’ copyright lawsuit against OpenAI, which alleges that the company trained its model on thousands of the paper’s stories without permission. The case is now in discovery, with more than 20 million ChatGPT user logs under review. This will set a price for content. It will not answer the underlying question, which is independence. That is being settled by checkbook before any judge rules. A captured press cannot hold power accountable. A synthetic press can do even less. Eight months from the midterms, the country is running both experiments at once. Those elected will eventually have to sort out the real fake news.

---

Reuben Steiger is a writer and entrepreneur based in Princeton, NJ. Over a 25-year career he has helped start companies including Second Life and has led global innovation for companies including Interpublic and Omnicom. His current focus is the scaling and adoption of AI technologies. He collects books about the future.