I have disliked books written by people. This is an awkward fact for the fashionable argument that a flat sentence must have emerged from a machine.
In October's new-release discussion on Reddit, some readers used comparisons with AI to describe prose they found disappointing. I am not repeating their allegations about particular authors. A reader's impression of a style does not establish how a manuscript was produced.
It does raise a question I keep coming back to in the secondhand shop: when did saying that a book bored us stop being enough?
At the shop we hear all kinds of strong opinions. Too slow. Too neat. Too many characters. Not enough happening. A novel someone abandoned is another person's favorite, which is an excellent arrangement for a shelf that depends on books moving between readers. The disagreement tells us something about taste and timing. It does not require a hidden production scandal.
I understand why people are wary of generated material being presented without clarity. Authorship is part of what some readers are choosing. If a publisher makes a claim about a work's process, a reader can reasonably want that claim to mean something. What I do not understand is the confidence with which a disappointing paragraph becomes evidence for an accusation.
The criticism gets weaker when the diagnosis gets louder
Suppose I read a fictional novel and find the dialogue stiff. I can describe the stiffness. Perhaps every person uses the same rhythm. Perhaps the conversation explains something neither speaker needs explained. Perhaps the prose skips the awkward silence that would have made the scene convincing.
That is criticism someone else can inspect. They can reread the scene, agree or disagree, and explain why. Announcing that a robot must have written it replaces the description with a claim about a process I did not observe.
It also misses the chance to discuss an ordinary weakness in a human book. People can write hurriedly. They can over-explain. A style that worked in one novel can become mannered in another. Editors and readers can differ about what should stay. None of those possibilities needs to become a certain diagnosis either, but they remind me that a poor result has more than one possible history.
My own reading notebook contains embarrassingly severe verdicts on books I later liked. I was tired. I wanted a different kind of story. I had not yet become interested in the thing the book was asking me to notice. The old verdict was an honest account of that reading. It was not the last possible account.
Leave room for a plain opinion
I can decide not to continue a book without proving misconduct. I can say a passage feels generic without identifying the instrument that produced it. I can ask a publisher for clarity without accusing a named writer on the strength of a hunch.
These distinctions make the conversation more interesting, because they leave the text available for discussion. Which image felt earned? Where did the dialogue lose a person? What changed between the books that made one feel alive and another feel mechanical? Those questions can teach me something about reading and writing.
The accusation mostly teaches me how satisfying an accusation can feel.
The shelf has room for human disappointment. It has room for books whose process is explained and readers who choose differently because of that explanation. What I would remove is the shortcut from I did not like this to I know how it was made.
A boring book can simply be a boring book. That conclusion has survived quite a long history of people writing them.
A thought worth keeping?
