The Future of Truth is a book about how AI is reshaping truth. It came out on May 12, 2026, with a foreword by Maria Ressa, the Nobel Peace Prize–winning journalist. Within a week of publication, The New York Times had found more than six quotes in it whose sources were fabricated or misattributed. One was pinned on the tech journalist Kara Swisher, who said flatly that she had never said any such thing. It was a sentence AI had invented.
The author, Steven Rosenbaum, explained that this was an accident, not deliberate fabrication. Early in the writing, he said, he had used AI as a search tool, flagging anything that came from it and sending it to the publisher's fact-checker. The fake quotes ran in the book anyway. A book meant to warn about the ways AI threatens truth had been brought down by exactly that mechanism.
It's easy to read this as the familiar warning that “AI makes things up.” But peel back a layer and a different structure comes into view. AI didn't create a new problem; it exposed a gap that publishing had left open for a long time.
A Structure Where No One Owns the Checking
In nonfiction publishing, fact-checking is not a contractual obligation. According to New York Magazine, which reported on the case, hiring an outside fact-checker costs anywhere from $7,000 to more than $10,000 per book. Because of that cost, the step has long been skipped. In the same report, a senior nonfiction editor at a major publisher said that after finding errors in a book and raising the question of fact-checking, they were told the publisher “has no responsibility to fact-check or to hire outside verifiers, and that doing so would shift the liability onto the publisher.”
Authors are unlikely to spend heavily on verification out of their own pockets, either. One literary agent said there is “no established system.” The issue comes up in every publishing contract, the agent said, yet no one ever resolves it. Few publishers have specific contract clauses governing AI use. One editor said they do send authors guidance not to use AI in certain ways, but that it “isn't legally binding, and once you're in the editing stage it ultimately comes down to trust.”
Responsibility for verification is pinned nowhere. The publisher says it has no obligation, the author can't absorb the cost, the agent says there's no system. Most of the time this gap never surfaced as a problem. At the speed a human writes, errors crept in only rarely, and that much was tolerated. When AI pushed production speed up, the unassigned responsibility turned into an accident.
The Forecast of an Age That Publishes Before It Verifies
This structure had already been diagnosed before AI arrived. In Too Big to Know, David Weinberger argued that the internet changed how knowledge works by eliminating the cost of publishing. A strategy of publishing first and verifying later took hold, he observed, producing a vast heap of data made public—and accessible to anyone—before it has been checked. The way facts once served as the foundation of knowledge changed in the process.
What Weinberger saw is that when the cost of publishing approaches zero, fact-checking is the first thing to be pushed aside. In the print era, the cost of printing and distribution itself was a barrier to publishing. Because producing a single copy cost money, publishers had to choose what to put out, and review entered that process. The internet erased that cost. Once publishing became free, the filtering step disappeared along with it.
AI went a step further. If the internet made publishing cheap, AI made production cheap too. Now that one person can generate a book-length draft in a matter of days, only fact-checking remains expensive and slow. That work is hard to automate, because a human has to cross-check every original source one by one, and the structure in which a single book runs into thousands of dollars hasn't changed. Production has sped up many times over while verification can't keep pace. The wider the gap between the two speeds, the greater the chance that unchecked sentences harden into a book.
Rosenbaum's book is an accident produced by that gap. AI generated the sentences fast, verification was pinned nowhere, and the result hardened into a book and went to market. Back when people wrote line by line, it would have been hard for six fake quotes to collect in a single volume. Once production speed overtakes verification speed, errors stop being rare mistakes and become a product of the structure itself.
The Solo Creator Has No One to Hand the Checking To
Here the position of the solo operator and creator is even starker than the publisher's. A publisher at least has several layers—editors, outside checkers, legal review. For the solo creator, all of those layers collapse into one person. There is no one to pass the verification to.
Someone who publishes their own e-books, writes their own newsletter, and produces their own video scripts controls every stage of production alone. Using AI makes that several times faster. Where it once took days to go from research to a first draft of a single piece, the process now shrinks to a few hours. The problem is that verification doesn't speed up to match. When you found and read the material yourself, you wrote with the sources in hand; when you write from AI's summaries, sentences that never passed through the original source go straight into the piece. The step a publisher skips because of cost, the solo creator skips for lack of time and effort. The result is the same: an unverified sentence goes to market under your own name.
Rosenbaum had a publisher and a distributor behind him, and still the responsibility was entirely his. In his statement he accepted full responsibility for the errors. The solo creator doesn't even have anyone to share it with. Even if a sentence was written by AI, the person who put their name on it is accountable for it. Putting your name on something means vouching not for where it came from, but for whether it's true.
So how far, and how, should you verify? Handing every sentence to an outside expert isn't realistic for a solo creator. Better to narrow the scope and concentrate.
Where to start? The point where Rosenbaum's accident began points to the answer. The fakery came from a direct quote that put words in Kara Swisher's mouth she had never spoken. Quotations are what AI invents most convincingly. It's safer to confirm against the original source that words inside quotation marks were actually said by that person. Next come numbers and dates. AI confidently offers specific figures like “about 70%” or “2023,” but that confidence is separate from accuracy. Statistics, amounts, and years have to be cross-checked against primary sources. Last are proper nouns. People's names, company names, book titles, the names of events—when AI gets these wrong, the sentence still reads smoothly, so they're easy to slip past. You have to stop and confirm they actually exist. Quotations, numbers, proper nouns: these three are where AI's errors leak out most often, and where a human can most reliably block them before publishing.
There's a reason to focus on these three. All of them are items whose truth resolves one way, so right and wrong can be checked independently. Interpretation and opinion, by contrast, can't be settled by cross-checking against a source. For a solo creator whose hands can't multiply without limit, it's reasonable to pour your time into the items that can actually be checked. Flag the information you took from AI, and build into your workflow a step that revisits only the flagged sentences before publishing. Rosenbaum did go as far as flagging the AI-derived information. He just left out the step of checking after flagging.
You Gained the Speed—Now Redesign the Checking
AI has given the solo creator production power on par with a whole publishing house. But it hasn't handed over the verification stages a publisher has along with it. Production can be delegated; the responsibility of putting your name on something cannot.
What Rosenbaum's case leaves behind is not a lesson to stop using AI. He said he used AI only as a search tool, but he never pinned the responsibility for verification on anyone. The more you hold a fast production tool in your hands, the more you have to decide first who verifies, when, and how. The gap a publisher leaves open because of cost, the solo creator can design back into their own work. If you can't place the verification stage outside, making the final pre-publication check a single fixed procedure is the realistic alternative.
The era of publishing before verifying has already arrived, just as Weinberger forecast. Kara Swisher learned that words she never said had been printed in a book only after the fact. If someone had cross-checked that one sentence against the original source before publication, the book wouldn't have collapsed. When you put your name on something and send it out into the world, deciding who does that cross-check, and when, has now become part of the work.
