The AI Con, Reviewed: Right About the Sales Pitch, Wrong About the Plan

Emily M. Bender and Alex Hanna spend 288 pages arguing that "AI" is a marketing term. I've spent most of my working life writing marketing terms, mostly for people who sell tires, tortillas and boat rentals, and on that point they're dead right. Easiest call in the book. Where they lose me is later, when the plan for fighting back turns out to be refuse the stuff and make fun of anyone who doesn't.

The book came out May 13, 2025. Fifteen months later it sits at 3.69 on Goodreads across 1,827 ratings and 376 written reviews, and it's still moving on Amazon, number six in Corporate Governance. That 3.69 is the number that matters here. A book everybody hated lands at 2.4. A book everybody loved lands at 4.4. A 3.69 with 663 four-star ratings on top of 502 three-star ratings means a large room of people who agree with the argument and still can't quite recommend the thing.

I know that feeling. I had it.

The core claim is that "AI" is a label salespeople hang on a pile of statistics so the pile sells at a premium. One Goodreads reader called it a book that exposes AI as "primarily a marketing term masking corporate power grabs, data exploitation, and the devaluation of human work." Fair reading. From where I sit, it isn't even controversial.

Here's the ground-level version. A vendor gets thirty minutes on a client call and runs a deck where "AI-powered" shows up on eleven of fourteen slides. The tool underneath resizes images and suggests hashtags. Useful. Also a script that ran fine in 2018 under the name "automation" and didn't command a per-seat license. Nothing about the software changed. The noun did, and the noun is worth money.

I've watched the same trick on the buying side. A brand manager at a CPG company hears a competitor is "using AI" for social and comes back wanting the same thing with no idea what the thing is. A line item appears on the scope. Everybody agrees it matters. Six months later nobody can point at a number that moved. Bender and Hanna are describing my inbox. Their publisher's copy promises a book about how AI hype "covers for a small set of power-hungry actors at work and in the world," and on that promise it delivers. I didn't need convincing. I've seen the invoices.

The middle of the ratings curve is where the honest criticism lives, and on this book the middle is huge. Twenty-seven percent of readers gave it three stars. Read those and the pattern shows up fast. They agree with the diagnosis. They can't stand the bedside manner.

"It's a very snarkastic book," wrote one reader at three stars, who went on to say that "There were times when I wasn't sure what the authors were trying to tell me, but I sure as hell knew they felt it was very smart." Another three-star review called it "mostly what other AI skeptics have written but containing less insightful analysis and more retellings of popular dissent," and said it would work better "as a limited podcast run." A third summed it up as a book that "had the essence of mocking something (AI hype) for the author's and audience's comfort."

That last line is the one I keep coming back to. It names the real business problem. Mockery is a closed loop. It rewards the people already in the room and gives everyone outside it nothing to do on Monday.

The two-star reviews cut sharper. One reader who wanted to love the book wrote that "the authors let their desire to deploy 'ridicule as praxis' run wild," the authors' own phrase turned back on them, and added flat that "The book constantly contradicts itself." Another two-star review landed the line I would have written if I'd gotten there first: "One is left with the feeling of reading a sermon, not an analysis."

The most useful negative review on the page comes from a reader who runs an AI consulting company. She says so up front. "I have a consulting company that helps businesses adopt AI," she writes, then explains she came in open, wanted the criticism, and couldn't finish, because "The tone of this book was so condescending and patronizing that I couldn't take it seriously." She is exactly the reader the authors most need. She is exactly the reader they lost. That isn't a small miss. That's the miss.

Doesn't make it a bad book. It's a good book with a bad ending, which is more frustrating than a bad book.

The first chapter alone is worth the cover price. A five-star reviewer with a technical background called it "one of the best plain English primers on neural networks, n-grams, and embeddings" they had read, and that matches what happens when I hand chapters to people who aren't technical. If you've ever tried to explain to a founder why the model doesn't know anything and can't be argued with, this book hands you the language.

The sharpest idea in it is one a five-star reader pulled clean: the boosters and the doomers are running the same play. "Those who boost the benefits of AI and those who see it as a potential source of our doom aren't actually on opposite sides in this debate," he wrote. Both camps need you to believe the machine is enormous. One sells salvation. One sells the apocalypse. Both quote the same inflated capability sheet. That reframing has changed how I read press releases. I don't say that about many books.

Even the annoyed readers hand it real credit. "I have SUCH mixed feelings about this book," wrote a four-star reviewer, who agreed with the central argument and then spent a paragraph on the politics, warning that "you may be left gagging on the extreme leftism." Another reader, no rating at all, called it "the passive aggressive Christmas present of 2025, folks," which is funny and roughly right about who it's for.

The prescription is where I get off. The book's answer to the con is stop using the tools and ridicule the people who keep using them. That's it. The authors reach for the Luddites as the model, and one three-star reviewer put his finger on the problem in nine words: "We should be AI Luddites. But the Luddites lost."

Another three-star review said the book "falls short in elucidating any solutions that have any chance of being enacted," and a reader who works with students wrote that "their proposed actions of don't use it and poke fun at it when others do, seems unsustainable to me." She's talking about teenagers. I'm talking about a food and beverage client with four SKUs launching in March and a photo budget cut in half. Neither of us gets to abstain as strategy.

I'll put my cards up, since this is a book about honesty in claims. My shop uses these tools every day. They cut transcripts. They log footage. They draft the first bad version of a caption so I have something to argue with. And yes, an AI system helped assemble the research behind this piece, which is exactly the disclosure the book would demand and rarely gets. What they don't do is any of the things the deck promised. They don't have taste. They can't tell you the tortilla in frame looks cold, or that the light on a can of beer at four in the afternoon in west Texas is going to blow out on the shoulder no matter what the histogram says. The judgment is still the job. The typing never was.

So the honest position, the one the book won't take, is that "AI" is an inflated marketing term and a set of genuinely useful production tools at the same time. The reason that sounds like a dodge is both halves are boring. Bender and Hanna are academics writing a polemic, and a polemic can't hold two things at once. That's why the last section reads like a rally and not like advice.

Buy it if you sit in a room where somebody keeps saying "AI" like it means one thing, and you need vocabulary to push back without sounding scared of computers. Chapter one alone will fix that.

Don't buy it expecting a plan. One reader compared the whole book to an episode of Last Week Tonight, informative on some things, deliberately vague on others, and satisfying mostly to people who already agreed. Fifteen months on, with the authors calling the models "these piles of racist algebra" and the market having spent that stretch making the bubble bigger rather than smaller, that comparison is holding up better than the prescription is.

The con is real. Bender and Hanna proved it. Somebody else is going to have to write the part about what you do on the Monday after you believe them.

Book: The AI Con: How to Fight Big Tech's Hype and Create the Future We Want, by Emily M. Bender and Alex Hanna. Harper, May 13, 2025, 288 pages. Goodreads: 3.69 across 1,827 ratings, https://www.goodreads.com/book/show/217432753-the-ai-con . Amazon: https://www.amazon.com/AI-Fight-Techs-Create-Future/dp/0063418568 . All reader quotes are short excerpts from public reviews on Goodreads, captured August 13, 2026.