Actually, It’s Ok to Steal Your Ideas. Sort Of… (Or: Learning to Love My Literary Influences)

Last month I found myself on a Zoom call with the members of a book club who’d read my debut novel; truly, a bucket list moment in my writing life. Near the end of the call, someone asked the question I suspect every fiction writer gets at some point: Where do you get ideas from?

Simple, I said: I steal them. I was joking. Sort of.

Creative types have been delivering versions of this quip for at least a hundred years. Circa 1920, T.S. Eliot wrote: “Immature poets imitate; mature poets steal.” Steve Jobs used to claim—without any real evidence—that Pablo Picasso liked to say: “Good artists copy, great artists steal.” And then there’s this line from Achtung Baby, which is permanently stuck in my head: “Every artist is a cannibal, every poet is a thief / All kill their inspiration and sing about their grief.”

I’ve been writing for twenty-five years, and I’ve produced more than a few essays about craft; but the whole creative process remains as obtuse as the inner workings of a large language model. By which I mean, even if you could map out precisely how you wrote a book, and who or what inspired you to write it, this explanation would be unlikely to satisfy anyone. Creativity is a black box that we use at our own risk.

A successful midlist writer once told me that when he got stuck writing, or found himself uninspired, he re-typed the words of writers he admired. This advice came in handy for me one afternoon while laboring at a novel. I retyped a scene from Mary Shelley, specifically the passage when Victor Frankenstein discovers the source of life itself, the spark that animates us all. To my surprise, the advice worked. I found that I was able to write my own scene with increased confidence.

There are some very obvious dangers for anyone who sources new ideas by sifting old ones. The Furies of plagiarism don’t bother to knock when they come for you. Every few years, a major writer or politician gets caught lifting passages, usually to traumatic effect. I’m naive enough to believe that many such incidents are accidental. It is possible to borrow without realizing; it’s even possible to echo something you’ve never heard.

My debut novel was still in an early draft when a friend read the manuscript and said it showed promise, but then asked coolly if I’d ever read Ursula Le Guin’s The Lathe of Heaven. Yes, I had read and admired books by Le Guin. But not that one. I was vaguely curious until I read a synopsis online; then, I was horrified. Le Guin in this book was mining the same ideas as I was. Her plot sounded a lot like mine. And she published her book five years before I was born.

I ordered the book online and waited impatiently for its arrival. No shipment has taken longer to reach me. I stumbled everywhere through a halo of dread; my most base writing fear had come true. Someone else had written the book I wanted to write; the original idea I thought I had was but a mere likeness of the true idea.

After the Le Guin novel arrived, I read it in less than a day. Almost from the first page, the anxiety eased; certainly, there were parallels, artistic coincidences. But as I rappelled deep into her story, I kept thinking, Ah, yes, this is similar but different, and by the time I finished, I saw my book not as an imitation of hers but as a new take on old questions of control, power and morality.

The absurdity of unintentional imitation is not limited to characters or plots. In real life I am often asked by strangers if I am aware that I look like someone else. I was ringed round at a friend’s birthday party last year by strangers who kept saying that I looked and acted just like their friend Chris. It was amusing, but annoying, after a minute or two. I am not Chris.

During my long hair and goatee days, strangers would tell me I looked like someone famous. The face they saw in mine was often an actor with a recent hit movie. Sometimes, a popular musician.

Now that I’m well into my forties, I am told I resemble a certain middle-aged action star with shaggy hair and a beard. The comparisons always make me uncomfortable. My wife and kids find this all deeply amusing.

Such comments don’t actually tell you anything about me. Something happens when people look at strangers and try to really see them—they grope for something familiar, something knowable. This guy with a beard kind of looks like that guy with a beard. This guy has blue eyes like that other guy has blue eyes. It must mean something, yes? It must be worth noting, right? Yet, the similarity is almost always a dead end.

Recently, I used a large language model to produce a list of novels published last year that had themes similar to the book I published. I was poking around, searching for comparables for a new book I was working on. The chatbot supplied a list of books, many of which I’ve read, but there was one I had not heard of before—Ghostlight by Clara Imogen. I asked for a summary of the book, and I got this:

This novel blends speculative fiction with literary depth, focusing on a world where people can communicate with “shades,” digital avatars of deceased loved ones. It examines grief, memory, and the limits of technology in replicating human presence.

Reading this synopsis, I felt sick and listless and dead all at once. Because it was happening again. This all sounded familiar. The book I am writing is not called Ghostlight—but it is called Ghost Machines. And the ghost machines in my novel are the digital avatars of lost loved ones. An idea that I thought was original to me suddenly wasn’t original at all.

I tried to take it in stride. I did not succeed. That halo of fear fell back upon me: that my idea was too similar to something else, that I would have to abandon all the recent work that I had done.

I searched the internet for more information on Clara Imogen. She had no other books that I could find. No Wikipedia entry. No announcement from when she sold Ghostlight—and that’s when I realized what was happening. She was an A.I. hallucination. So was her book.

I have never felt so relieved to identify misinformation. The large language model made up Ghostlight because such a book was plausible; because statistically, it presumed Clara Imogen was more likely to exist than not.

In the history of stealing ideas, no product has ever stolen as much as large language models, which feed on the entire internet in order to keep their digital hearts beating. Where do machines get their ideas from? From us. For as long as we have ideas left to give. Or old ideas to make new again.

In “Pierre Menard, Author of the Quixote,” Jorge Luis Borges presents the story of a contemporary author who writes Don Quixote. Not a new translation, not a modern retelling—through his own laborious process Pierre Menard writes the classic novel using the exact same words as Cervantes. Yet this new Quixote—according to his readers—is utterly changed because of the context of its creator.

In this story Borges is playing with the reader, presenting a farce as if it really could happen. Or is he? Is Menard’s claim really all that different from how the same person can seem different in different contexts? Can the same words have different meanings when spoken on the job, or at a party; by a person, by a bot?

Of course. Everything depends on context.

We are always interpreting one another, whether we’re reading books or chatting at parties or staring out the window of a moving train. To understand anything, to really grapple with it, requires a person to place people, ideas, characters, even sentences in a larger context of meaning.

Artists exist in that context, just like everyone else; our best hope is to use the familiar to say something new, something meaningful, in spite of itself.