Don’t Join the Zombie Army!

-first published in Substack blog “The “Last Editor” on November 19, 2025

I don’t know if you’ve ever read the 2006 novel from Max Brooks, World War Z: An Oral History of the Zombie War. Even if you’re not a fan of science fiction or horror, the book—and I’m talking about the book version and not the 2013 Brad Pitt (M-I-Z!) film which is just sort of meh—is an amazingly journalistic telling of what a zombie apocalypse might be like. The protagonist of the novel, a fictional version of Brooks himself, conducts a series of interviews to write the definitive account of humankind’s war with the walking dead. The novel is a series of these interviews, deftly delivering the exposition of what happened in the war, all while giving us characters about whom we can care.

One particularly striking point of the book comes as the fictional Brooks interviews General Travis D’Ambrosia, the Chairman of the Joint Chiefs of Staff. Gen. D’Ambrosia describes the challenges of fighting an enemy—the zombies are nicknamed “Zach” by the military—with one particular strength on its side:

“…eventually you’re going to hit the ceiling…unless every time you killed an enemy, he came back to life on your side. That’s how Zack operated, swelling his ranks by thinning ours! And it only worked one way. Infect a human, he becomes a zombie. Kill a zombie, he becomes a corpse. We could only get weaker, while they might actually get stronger.” (Brooks, Max. World War Z: An Oral History of the Zombie War, p. 272).

This is one of the most eye-opening notions in the book. In normal war, when one of your soldiers is killed, that’s the end—that soldier is off the battlefield. But in the zombie war, your dead soldier gets up and joins the enemy!

While zombie wars are still—at least to this point—a work of fiction, we have a situation much like Brooks describes happening to journalism.

Our profession has long been one that chews people up and spits them out. The notion that someone would graduate with a journalism degree and spend the next 45 years or so working in newsrooms is almost as much a piece of science fiction as is zombie literature. Almost no one lasts that long. I’ve long had the opportunity to watch the trajectory of young journalists. Once they graduate from our program at Mizzou, it’s easy for me to follow their careers. Nearly everyone lasts for a single contract as a reporter, producer or in some other newsroom position. After that first contract is up, a number of them call it quits. That pattern repeats with a smaller cohort sticking it out after each successive contract. And the rate of attrition is rising as we lose more people sooner. There are plenty of reasons for this rising rate—more than I’ll go into here today. What I want to address is what happens when they leave journalism. Some of them are joining the side of the zombies.

In the Brooks novel, new ranks to the zombie corps join automatically upon death. Ours in the real world do so with a bit of coaxing.

I’m talking about the current campaign to hire journalists to train AI—mainly Large Language Models (LLMs) and machine learning—to do the work of reporting news stories. Companies like Mercor have ongoing outreach to get journalists to join teams that will train AI on how to do the job of a journalist. Here’s Mercor’s pitch:

“Mercor is seeking experienced News Analysts, Reporters, and Journalists to support a leading AI lab in advancing research and infrastructure for next-generation machine learning systems. This engagement focuses on diagnosing and solving real issues in your domain. It’s an opportunity to contribute your expertise to cutting-edge AI research while working independently and remotely on your own schedule.”

Seems harmless enough on first reading, right? Of course, it’s not. This notion of “solving real issues in your domain” refers to the real issue that AI cannot do what a good journalist can do. That becomes obvious when you read the rest of the ad. For key responsibilities, it says you’ll be “asked to create deliverables regarding common requests within your professional domain.” Those common requests? I’m guessing they are something along the lines of: “How can I lay off all my human reporters and get AI to put out a newspaper/newscast for me?”

To add insult to injury, the Mercor position pays $60 to $75 an hour—far more than the journalists it’s recruiting (“4+ years experience”) would likely make in their journalism jobs. To be clear, these aren’t full-time positions. But that sort of pay would be attractive to people who have left the business and want to cash in on the journalism skills they are no longer using.

While I’ve focused to this point on people leaving our business, I know some currently working in journalism might be attracted to earning $75 an hour to do this work on the side of their current duties. The argument against doing that should be even stronger for them. They’re still working in the profession and could have a long career ahead of them. But by training AI, they would be actively working to eliminate jobs that they themselves could get some day.

Am I completely opposed to AI? No—not even using it for journalism.

Some newsrooms are using AI to manage the deluge of news releases they get each day, picking out the ones that deserve attention from all the chaff. Others are using AI to search through public meeting agendas to look for stories worth sending a human reporter to cover. And I, for one, never want to transcribe an hour-long documentary interview by hand again. I’m happy to let AI do that. But I do not want to watch a twelve-hour documentary on the American Revolution made by AI. I still trust Ken Burns to provide me the best version of that sort of journalistic art.

And art is really what we’re talking about here. Good journalism combines a great deal of science with a little bit of art. The science comes from the logistics of getting to all the right places to get footage and do interviews, from the mechanics of writing a good sentence and the from the inarguable visual logic of a wide-medium-tight matched action sequence. But the art is just as important, coming from coaxing an important character into telling her story, from turning just the right phrase to marry words, pictures and natural sound and from the passion to keep pushing when everyone is closing doors in your face. AI can’t do art. It may try to imitate it, but it will never achieve it.

I’m half tempted to join one of these AI training efforts to infiltrate the enemy to see what it’s up to. But I don’t want to end up listed as a collaborator. It would be interesting to see if we could sabotage the effort. It would take a concerted effort to get in there to mess with the machine. Let me think about that one…

World War Z ends (spoilers ahead) with the humans coming out on top, beating back the zombies into some isolated areas and holding them at bay in a world badly battered by a costly conflict. I’m not sure if I can be as optimistic about our own zombie war. There are many collaborators on our side helping the enemy gain strength and figure out its strategy to win. All I can ask is that the people I know not be one of them.