Why a non-techie wrote one of the most useful AI guides for newsrooms

Martin Schori spent three years building one of Europe's most closely watched newsroom AI operations, the AI hub at Sweden's Aftonbladet. (Here’s my case study from February). Schori has now left Aftonbladet and wrote a book about what he learned. It’s called “AI in the Newsroom: Everything You Need to Become Tomorrow’s Journalist.” This guide was published first in Swedish and is now available in English. I read it, questioned him about it and wrote a review. As a News Machines subscriber, you get a discount (free international shipping) when you order Martin Schori’s book. Paying subscribers can also download one of the core chapters: Chapter 7: Journalist 2.0: The Human at the Center When AI Takes Over the Boring Jobs. (The discount link and the free chapter download link are at the end of this post).

Martin Schori is not interested in technology for its own sake, and he wants you to know it. He doesn’t follow the field closely and doesn’t read a lot of newsletters about AI. He came to AI as a journalist trying to solve journalism’s problems, not as someone drawn to the technology. This isn’t a case for staying ignorant of how the tools work; the book includes a glossary and a plain-language explainer of what a large language model is actually doing when it answers. It’s a statement about altitude. Schori gives you the general sense you need to judge when to trust the output and how to use it — and then stops, because the mechanics beyond that aren’t his subject. His book is written from inside the AI machine room by someone who waded in reluctantly and took first-hand notes.

What you get, as a result, is a book with almost no daylight between its claims and its evidence. Nearly everything in it traces back to something Schori watched happen at Aftonbladet. When he argues that AI projects have to be anchored in the newsroom rather than handed to the tech department, it’s because he tried it the other way. When he says article summaries can keep readers on the page longer rather than shorter, it’s because the data surprised him. The book takes its authority from someone who has already made the mistakes so you can avoid them.

That posture extends to how the AI hub was staffed — four of its seven seats went to journalists, deliberately, and to journalists with credibility on the newsroom floor. Schori’s reasoning doubles as the book’s design principle: journalists won’t listen to tech people, so the case for AI has to come from one of their own, in their own language. The book is that case, made at length.

Most AI-and-journalism writing either sells transformation or sounds an alarm; Schori does neither. He tells an anxious newsroom the shift is survivable, and hands it somewhere to start. What he won’t do is pretend to know how it ends. Where does the media industry go once the tech giants control distribution? He doesn’t pretend to know. He provides a guide to the first moves, by someone careful about the distinction between what he’s seen and what he’s only guessing.

Schori’s book is for journalists on all rungs of the career ladder, but that doesn’t mean it’s for everyone. What unites the target audience is a starting point, not a job title — people who are not already steeped in the subject but who want to know more about AI. This is a book for the curious beginner, whether that beginner is twenty-three and interning or fifty-five and running a newsroom. Schori’s own proof that the audience runs older than just “young journalists”: a progressive Swedish publishing house bought copies for its entire staff — confirmation, he reads it, that even three years in, there are a lot of people in the media industry whom the early adopters tend to forget.

And who should skip it? People already ahead of the curve. If you’ve been building your own workflows for a year or longer and following the field closely, you’ll find little here you don’t already know. That’s not a flaw; it’s the book doing its job for someone else. If you’ve been reading News Machines for a while, you might not be the audience — but you probably know people in your own newsrooms who are, and for whom a calm, first-hand guide is exactly the push they’ve been waiting for.

The Aftonbladet AI hub was set up as a cross-functional team working on top of the existing organization, so it could move faster than the normal product-and-tech pipeline allowed. Its early output included an “AI buffet” — a bundle of editorial tools for SEO headlines, fact boxes, summaries, follow-up suggestions — meant to lower the barrier to trying AI at all. Some of it surprised the newsroom: readers who opened AI-generated summaries stayed longer, not shorter, which turned skeptical reporters around.

These experiments are also where the book already shows signs of aging. They are only three years old and already read like reports from an earlier era — which is the risk Schori took on knowingly. He half-jokes that the fastest way to feel irrelevant is to write a book about AI, and he wrote one anyway, betting the principles outlast the examples. It is the fair caveat to carry into the chapters that lean hardest on specific projects.

The AI buffet’s bigger failure was structural. The tools lived outside the content-management system, so using them meant extra clicks, logins, and remembering they existed — and in a newsroom built for speed, that friction is where adoption dies. Schori is clear-eyed about it. It is also the first thread of a larger doubt: the tools built to make the newsroom more efficient were the ones nobody could quite fit into the day.

The book teaches the first steps towards efficiency — summaries, transcription, translation, fact boxes. But talking to Schori now, with the hub behind him, you hear someone who suspects the whole sector aimed those first steps in the wrong direction.

The industry, he argues, made a definitional mistake early on: it treated AI mainly as a tool for efficiency, a coming productivity revolution. And because the newsroom is a media company’s biggest headcount, that’s where everyone went hunting for savings. The better question, he now thinks, runs the other way — not where can we save time and money, but where can we create the most value, which may mean starting outside the newsroom entirely, in subscriptions, marketing, or churn prediction.

The evidence is his own. The efficiency revolution at Aftonbladet and elsewhere mostly hasn’t arrived; some experiments cost more time than they saved — and the buffet that wouldn’t fit into the CMS was an early sign of it. The AI-powered workflows Aftonbladet spent heavily to build, he notes, can now be bought cheaply off the shelf. The payoff for publishers, in other words, is turning out to be about value capture, not cost-cutting — and the two demand different strategies.

So what does an AI team tell a Chief Financial Officer who was promised savings that never came? Schori’s answer is to change the question — to stop hunting efficiency in every corner of the business and ask where real value sits instead. But he adds a structural caveat familiar to anyone in one of these roles: the new AI jobs often have no mandate, and it is very hard to drive transformation without a say in strategy. Win that seat and the CFO conversation happens between equals; without it, the AI team becomes a service organization, taking orders for tools nobody strategically chose.

For a small newsroom — one without Schibsted’s resources — Schori’s advice is more startling: it may be wise to do very little for now. The urgency that says act immediately or lose your business within months is, he suspects, overblown. The tools are getting cheaper and better on their own, so a smaller newsroom can reasonably buy what it needs later, or simply wait. Coming from the man who ran one of the most ambitious newsroom AI programs in Europe, that is a striking thing to hear — though he means it for the publishers least able to afford a wrong bet, not as a general license for paralysis.

And there is a corollary to “stop chasing efficiency.” If the point is not to do the same work with fewer people, then the time AI frees up is only worth freeing if it is spent on the work that still needs a human — the reporting, the judgment, the relationships. Saved hours are not the return. What you do with them is.

Read for what it is, Schori’s book delivers: a calm, practical first guide for the many people who haven’t truly started yet building AI into their workflow, written by someone who has — and who is honest about how fast AI experiments date. The sharper lesson comes from Schori himself, a step ahead of his own pages: the industry spent its first three years chasing efficiency in the newsroom, and the value was somewhere else all along. Schori’s book gets you moving. His insights beyond the book tell you which direction.

Here’s the link to the discounted book (= free international shipping).

Log in or become a premium subscriber of News Machines to download Chapter 7: Journalist 2.0: The Human at the Center When AI Takes Over the Boring Jobs.