What a piece of work wrought by the developers! How capable of reason, how infinite in faculty — generating poems, images, essays, and articles in a flash! How like an angel, answering every question, following every instruction, readily.
The answer to the prayer of the blocked writer, the unprepared student, the curious mind, these seemingly omniscient bots of artificial intelligence inspire riffs on Hamlet’s famous speech about “what a piece of work is a man”. But any admiration for them should be tempered with apprehension: they are a mixed bag, not an unmixed blessing.
Their powers of invention and detection are turning academia, media, and the literary world into police states. The self-assurance with which Shakespeare claimed his poetry could make his friend immortal (in Sonnet 18) stands in sharp contrast to the current plight of writers, subjected to the indignity of competition with chatbots and the scrutiny of AI detectors. Even young people with no masochistic desire to become writers share their pain — for everyone must go to school and write.
How widespread the pain has become in America is reflected in a recent Wall Street Journal report headlined “Students turn to AI to prove bots didn’t write their essays”. According to the report, teachers employ software to detect AI-generated content in students’ work, while students use the same tools to ensure their work appears entirely their own, untouched by bots. A student who insists he never uses AI to write essays nonetheless puts them through a detector — just to make sure they read like his own composition.
The irony is that students with a large vocabulary and a formal writing style are more likely to be suspected of using AI. Suspicions may be raised by words like “delve” and “tapestry,” or phrases such as “crucial role” and “valuable insights”. “It’s like you get penalised for being a proper writer,” Devan Leos, co-founder of Undetectable AI, told the Journal.
Adding to the confusion, different AI detectors yield different results — what’s flagged by one may be cleared by another. Consequently, some students use multiple detectors to ensure their work passes scrutiny on all fronts.
Google demands
This policing of AI in schools is matched by the same sharp oversight in the media. Online content writers may be required to run their pieces through the same detectors as students — or have editors do so — to confirm their work is entirely their own, 100% human, without the slightest input from bots. This effort to guarantee originality is admirable — and necessary, too, for that’s what Google demands.
Webmasters comply with Google’s preferences because the search giant determines what shows up on your screen if, like the vast majority of internet users, you use Google.
Google is right, of course, to prioritise original content. That’s what people want: something new and fresh. And Google has made its business to give them just that and satisfy their curiosity.
Google’s effort to prioritise originality and demote plagiarism is a service to the online community. No one wants to wade through junk to find useful information.
But where does literary borrowing end and plagiarism begin? A simple copy-and-paste from another article is undoubtedly brazen plagiarism. But literature and music have a tradition of borrowing from the past to create something new. Anyone who studied Shakespeare in school will recall annotated editions explaining the words, events, and sources he drew on to conjure something uniquely his own.
“I love to lose myself in other men’s minds,” confessed the essayist Charles Lamb. How could a writer like him not echo the words of others?
T.S. Eliot quotes Dante, Spenser, and the Upanishads in The Waste Land. One may argue these are references and allusions, not plagiarism. True. But will the current hullabaloo over plagiarism make writers wary of borrowing in order to create? One hopes not—for literature thrives on recreation.