Are we all plagiarists now?

It is a truth universally acknowledged that people get away with plagiarism a lot—even when the line they are plagiarising is “It is a truth universally acknowledged…” In 2007 chapters of Jane Austen’s “Pride and Prejudice” were sent with minor variations to 18 editors, purporting to be a “new” novel entitled “First Impressions”. Despite it containing one of the most famous lines in English literature, only one editor called out the hoaxer. “My first impression” on reading “First Impressions”, he replied with Austenian archness, was “mild annoyance” then “laughter”.

Many find plagiarism less droll, as a new book by Roger Kreuz, an academic, makes clear. His book canters through 20 centuries of plagiarism, from musical plagiarism (Bob Dylan) to literary plagiarism (Dylan Thomas), oratorical plagiarism (Joe Biden) and all-of-the-above plagiarism. Mr Dylan’s acceptance speech for the Nobel prize in literature, which referred to “Moby Dick”, had echoes of a less-than-noble literary source—SparkNotes.

This book, which offers more anecdotes about plagiarism than an argument about it, feels like an opportunity lost. Plagiarism is an old vice—the word was borrowed (plagiarised?) from a poem by Martial, a Roman writer—that feels newly relevant. What counts as intellectual theft—and what is considered acceptable borrowing or inspiration—are the great questions of the AI era. That is true both personally (is it alright to use AI to write that job application or love letter?) and legally (will popular AI models be punished for training on copyrighted material?).

Part of the problem is that there is no precise definition of plagiarism—it ranges from verbatim copying to the fudgier theft of concepts. As with pornography, it is assumed you “know it when [you] see it”. And you see it a lot now, thanks to technology. When Shakespeare nabbed the barge scene in “Antony and Cleopatra” from Plutarch, he had to write it out by hand, which was a bother. To churn out copies of knock-offs like “Martin Guzzlewit” and “Oliver Twiss” took Dickens’s imitators time and typesetting. Since the advent of the “Ctrl” and “C” keys, plagiarism has become easier to do—and harder to Ctrl.

It is even harder now. Large language models (LLMs) like ChatGPT and Anthropic’s Claude are trained on vast quantities of written, often copyrighted, material, so some feel that their output rips off authors and other creative types. Writers have filed lawsuits against AI companies. In one complaint, plaintiffs said tech firms were committing “systematic theft on a mass scale”. Anthropic recently agreed to pay authors $1.5bn for using 7m pirated books for training. (This reviewer could receive a payout from that settlement for one of her books.) On January 15th two publishers asked to join a class-action lawsuit against GoogleAI for engaging “in one of the most prolific infringements of copyrighted materials in history”.

What about users of these LLMs? Training oneself on millions of books, then spitting out your own words afterwards, is a very good definition of what a writer does. Moreover, using ChatGPT, Mr Kreuz argues, does not make you a plagiarist, since it is not cribbing from a single “original” text. He suggests LLMs are doing unacknowledged “ghostwriting”. To many that is too generous: this is still plagiarism, but with an AI accomplice. It is a layered larceny, almost a double crime: the AI steals and often repeats the words it was trained on, then you take those words, as your own, for profit—the profit of imposture.

It is a thorny issue. But then it always was. Creativity and originality have been in a difficult dance for centuries. A list of authors accused of filching is a “Who’s Who” of literature: Chaucer, Shakespeare, Swift. Some, it is true, did take a principled stand: 250 years ago Laurence Sterne inveighed against plagiarists who, like apothecaries, merely pour content from “one vessel into another”. As it turned out, he had nicked that line from a fellow author, Robert Burton. There is, said Mark Twain, not “much of anything in any human utterance…except plagiarism!”

It is often said to be romanticism and its obsession with “originality” that caused people to care. But plagiarism has always irked. The Latin word plagiarius means “kidnapper”, a criminal who should, according to Roman law, be “thrown to the wild beasts at the first public spectacle”. The authors who brought a case against Anthropic might feel likewise.

And there is a murky but crucial distinction between inspiration and straight-up cribbing. T.S. Eliot said that “Immature poets imitate; mature poets steal.” Few ever bother to include the second part of the quote, which notes that “Bad poets deface what they take, and good poets make it into something better, or at least something different.” Any author takes at least some inspiration from other works. The case against Anthropic partially failed: a judge likened Anthropic to a “reader aspiring to be a writer” who uses others’ words to “create something different”.

What really changed things, Mr Kreuz argues, was less a new literary category than a legal one. With the creation of copyright in the 18th century, copying rights could be more precisely articulated—and copying wrongs more clearly prosecuted. Authorship, says Robert Douglas-Fairhurst, a professor of English at Oxford, “is a profession” so “you have to guard your property”. Charles Dickens sued printers who churned out knock-offs partly from authorial umbrage but also for money.

Technology changed things too: printing meant that more people could read and buy books—and check what was stolen. It is changing things again. Universities are increasingly turning to AI to spot AI-written work (even as students use services like Dumb it Down to make their AI-fuelled work sound more believable). It can be detected. Chris Caren, the boss of Turnitin, a popular plagiarism detector, describes plagiarised prose as “beige”: “well-written, but not very dynamic”. It has verbal tics: it is keen on dreary words like “holistic” and notably keen on “notably”.

The dance is changing again. And now it is not just authors who have to think about it, but all of us. Mr Kreuz notes that mentions of plagiarism in the New York Times have increased eight-fold since the 1950s. Perhaps because there is more. But also, surely, because it is fun. Plagiarism stories follow a perfect tragic arc: a flawed hero suffers a whopping comeuppance. It is a wonderful spectacle. And for what do we live, but to make sport for our neighbours, and laugh at them in our turn? Or did someone else say that? ■