Your Next Favorite Book Was Chosen by a Machine — And That Should Worry You

There's a version of this story where the algorithm is your best friend. It knows you loved The Midnight Library, so it serves up five more emotionally redemptive novels about second chances. You finish all five. You feel like a reader. You feel productive. You feel seen.

Here's the version nobody's pitching you: you just spent six months inside a very comfortable literary bubble that a machine built for you — and you never even noticed the walls.

The Promise Was Efficiency. The Delivery Is a Loop.

When Amazon first started recommending books based on purchase history, it felt like a revelation. Finally, a way to cut through the noise. Goodreads followed with shelves, ratings, and friend activity. Then TikTok arrived and #BookTok exploded into a full-on cultural phenomenon, moving millions of units of titles that a traditional publisher's marketing budget couldn't have touched.

On the surface, all of this looks like democratization. And in some ways, it is. Authors from outside the traditional publishing establishment have found genuine audiences. Backlist titles — books that had been quietly gathering dust for years — suddenly got second lives because an algorithm decided they matched a trending emotional vibe.

But zoom out a little, and a different pattern starts to emerge.

A 2023 analysis of Goodreads rating clusters found that the platform's recommendation engine disproportionately surfaces books that already have high engagement — which, unsurprisingly, are books that already have high engagement. It's a feedback loop dressed up as discovery. The titles that get recommended are the ones that get read, which makes them the ones that get recommended. Debut authors and literary fiction from underrepresented voices get filtered out not because readers wouldn't like them, but because the machine hasn't seen enough signal yet.

BookTok Isn't a Bookstore. It's a Vibe Engine.

TikTok's algorithm doesn't care what you read last year. It cares what made you stop scrolling three seconds ago. That distinction matters more than most people realize.

When a book goes viral on BookTok, it's usually because it triggered a specific emotional response — a crying scene, a plot twist, a romance that made someone visibly gasp on camera. Those are legitimate reasons to love a book. But they're also a very narrow slice of what literature can do.

Books that are slow burns, structurally experimental, or thematically challenging tend to be hard to capture in a fifteen-second video. So they don't get made into fifteen-second videos. So they don't get recommended. So they don't get read — at least not by the millions of people whose reading lists are now quietly being curated by an engagement-optimization engine.

Publishing insiders have started noticing the downstream effects. Several acquisitions editors have noted, off the record, that the pressure to pitch books as "BookTok-able" is real and growing. If a manuscript doesn't have a hook that translates to short-form video, it's a harder sell. That's not the algorithm's fault, exactly. But it is the algorithm's consequence.

Efficiency Isn't the Same as Depth

This is where the productivity angle gets genuinely interesting — and a little uncomfortable.

We're living in a moment obsessed with reading optimization. Track your pages. Hit your Goodreads goal. Consume more, faster, smarter. The algorithm fits perfectly into that mindset because it eliminates the friction of discovery. You don't have to wander through a bookstore or take a chance on a stranger's recommendation anymore. The machine handles it.

But some of the most meaningful books people have ever read came from friction. From picking up something weird at a library sale. From a professor assigning something you would never have chosen yourself. From a friend who knew you well enough to say, "trust me, this isn't what you usually read, but read it anyway."

Algorithms are genuinely bad at that last part. They optimize for preference matching, not preference expansion. They're designed to reduce the chance you'll bounce — not to introduce you to something that might change how you think.

What the Data on Reading Diversity Actually Shows

A study published by the Panorama Project tracked reading patterns across US school districts and found that students' independent reading choices were becoming increasingly homogenous over time — clustering around a smaller set of titles and genres. The researchers pointed to recommendation systems as one contributing factor, alongside social media influence.

Among adult readers, the picture is similarly mixed. Surveys from Pew Research consistently show that Americans who read regularly are reading more books than they did a decade ago — a genuine win. But the variety of what they're reading, measured by genre spread and author diversity, has narrowed in the same period.

More reading, less range. That's the trade-off nobody advertised.

So What Do You Actually Do With This?

The honest answer is that you don't need to swear off Goodreads or delete TikTok. Recommendation engines aren't evil. They're just limited — and the problem is that we've started treating their limits as features.

A few things worth building into your reading life:

Introduce deliberate randomness. Browse a physical bookstore or library with no agenda. Pick up something in a section you never visit. The friction is the point.

Follow humans, not algorithms. A thoughtful book newsletter from a single opinionated critic will expand your reading world faster than any recommendation engine, because humans recommend things that surprised them — not just things that matched a preference profile.

Audit your last ten books. If they're all the same genre, same emotional register, or all came from the same platform, that's useful information. Not a moral failing — just a data point worth knowing.

The algorithm isn't trying to make you a worse reader. It's trying to keep you engaged. Those two things just aren't always the same goal.

At House of Summary, we believe the best reading life is a curated one — but curated by you, not by a machine that's never actually read a book.