Could the next great novel be written by AI (and would you even be able to tell)?

Illustration: Pete Reynolds/The Guardian

Illustration: Pete Reynolds/The Guardian

Could the next great novel be written by AI (and would you even be able to tell)?

As allegations of LLM use rock the literary and media worlds, linguists explain what really distinguishes human and machine language, while novelists including Jennifer Egan and Jeanette Winterson reflect on the future of fiction in an age of ChatGPT

Three paragraphs, from three different hotel reviews. Can you tell which, if any, were AI‑generated?

“The hotel is in a great location for everything. Lots of places to eat and drink. The hotel itself is always abuzz. The tavern located on the ground floor is definitely a must. Food, service, prices and atmosphere were great.”

“A good hotel, though the room had the proportions of a well-appointed lift. Slept well, shower was excellent, staff were friendly. Breakfast was busy but competent. Would return, though probably not with a very large suitcase.”

“Excellent base for a London trip. The room was quiet, the bed comfortable, and everything worked exactly as it should. Staff were helpful without hovering. A smooth, unfussy stay from start to finish.”

How do you reckon you did? Most people, says Claire Hardaker, a professor of forensic linguistics at the University of Lancaster, get this kind of judgment right only about 60% of the time. Her online test, Bot or Not, asks users to identify the fakes in a series of 15 reviews. The middling success rate might come as a surprise to those convinced they can spot AI writing at 50 paces. When doubts were raised in May about the authenticity of a prizewinning short story by Jamir Nazir, social media users were lightning-quick in their condemnation. “If you know, you know,” commented one.

Hardaker says her respondents tend to rely on a few quick rules of thumb to identify AI language, including the presence of cliches and the use of dashes. The “rule of three”, where words or phrases are arranged in a satisfying trio, is also thought to be a giveaway. “People have learned very simplistic rubrics and now just madly apply them everywhere.”

There’s a problem, though: these “tells” are also characteristic of human writing, which, after all, the large language models (LLMs) that produce them were trained on. “You could go back to Charles Dickens and say he had AI, because he used the em dash too.” And orators have known about the rule of three ever since Julius Caesar said Veni, vidi, vici. In our hotel review examples, only the first one was authentic. Did you clock it?

Perhaps because it is so hard to know for sure, suspicion has become the order of the day. In the literary world, accusations of AI use now bedevil writers, with varying levels of justification. A debut horror novel, Shy Girl, was withdrawn by publishers Hachette after rumours circulated online that the author had relied on AI, which she denies; Steven Rosenbaum’s book The Future of Truth, a serious study of “how AI reshapes reality”, was found to contain numerous hallucinated quotations, which the author acknowledged in an apology.

Media organisations, including the Guardian, field increasing numbers of complaints about supposedly AI-generated text. These include intuitions about particular turns of phrase, but also comments about typos and grammatical errors. In one case, the word “after” was inadvertently duplicated in a sentence. “I can’t imagine a human editor/proofreader missing something like this,” wrote one reader, displaying a touching faith in our copy-editing abilities.

The problem is that not only does AI train on human writing, but humans are stylistically influenced by AI, the interplay creating a kind of linguistic hall of mirrors. Short of an author admitting it, it’s hard to say for certain whether an individual piece of writing is AI or not. That uncertainty is a recipe for paranoia.

And if you’re tempted to reach for a commercial screening tool to sort human from machine, that comes with uncertainty too, says Hardaker. “Given that some of us naturally write in a way that would be seen as AI-like” – she mentions neurodivergent people, for example – “that will be detected as AI. And you can modify AI output to make it seem more human-like. You put that kind of content into an AI detector, you’re going to get wacky results.” As someone who has served as an expert witness in court, she’s “extremely sceptical” about their efficacy.

The newly popular detector Pangram, which boasts false positive rates of around 1 in 10,000, has been shown in independent tests to be highly effective at detecting AI writing even when it’s been run through a “humanizer” app to disguise its origin. But questions remain. I was able to fool it on the first attempt (see the screenshot below) by channelling a bombastic register that might well be characteristic of AI, but could equally be the work of someone with a naturally bombastic style – or, more to the point, a writer who has been steeped in the output of the LLMs that power ChatGPT, Claude and Gemini. That, increasingly, is all of us.

Photograph: David Shariatmadari

Vast amounts of AI writing are now being published every day – from advertising copy to academic abstracts and fiction. At the same time, it looms ever larger over our lives via auto-generated email suggestions, “AI overview” search results, and the responses to our chatbot queries. At this level of exposure, it’s no longer a question of whether AI is changing language, both the way we speak and the way we write; the question is how. And should we resist, or embrace it?

We’ve known for some time that LLMs generate text that can be slightly different from human writing, on average. Often this only becomes clear when you look at large amounts of material. One eagle-eyed researcher linked the sudden popularity of the word “delve” to LLMs back in 2024 after searching a database of scientific papers. Other “focal words” that AIs have tended to overuse include “showcase”, “boast”, “underscore”, “garner”, “align”, “surpass” and “intricate”. But, again, any individual piece of writing could entirely innocently make use of this vocabulary.

In a further twist, some researchers think the “delve” phenomenon might not be down to the models themselves, but the humans tasked with evaluating and steering them in a process known as “reinforcement learning with human feedback”. For workers who are “underpaid, stressed, and under time pressure”, it seems “certain words are treated as a proxy for quality” and the model is inadvertently trained to use them more often. In other words, “delve” might owe its meteoric rise to the fact it doesn’t seem like the kind of word an AI would use. (A separate suggestion that it appeared more often because it was characteristic of English used in Nigeria, where many RLHF workers lived, isn’t borne out by the data.)

There are other patterns we can distinguish: LLMs love nouns, but they seem to use pronouns less than humans. This might reflect the fact they don’t do as much talking about themselves or other people as we social creatures do. They like attributive adjectives (“the uncomfortable chair”), but not predicative ones (“the chair was uncomfortable”), perhaps because they prefer to deliver information in small, dense packages, whereas we pad things out. Different models have clear idiosyncrasies – you might even call them “dialects”: Gemini enjoys saying “here’s a breakdown”, while Deepseek often responds with a cheerful “Certainly!”. When asked to edit formal English from around the world, AI tends to flatten and homogenise towards an Anglo-American standard, in a process researchers have termed “cultural ghosting”. Thus the perfectly acceptable request in Indian professional English to “Kindly do the needful & revert back at the earliest” gets “corrected” to “Please complete the task & respond promptly.”

The evidence that aspects of LLM-speak have escaped into the “real” world, changing the way humans use language when AIs aren’t around, is now rolling in. One study analysed thousands of unscripted conversations and found that words like “delve” and “boast” spiked after ChatGPT was released. Another showed the frequency of “delve” in academic abstracts actually dropping after it was singled out on social media, in a sign that AI’s influence might play out in complex ways.

Illustration: Pete Reynolds/The Guardian

Does any of this matter? Language changes all the time – words come in and out of fashion, and new technology has always been one of the forces behind this. AI does seem to be generating particularly high levels of anxiety, though. Why? “I think where it scares people is that idea of encroaching into sentience, of becoming the new human,” Hardaker says. Since 2023, she’s expanded the Bot or Not project into speech and music, and has noticed just how viscerally people react when a song they’ve enjoyed turns out to have been composed and performed by a machine.

Gary Shteyngart, a novelist who teaches creative writing at Columbia University, noticed a similar strength of feeling among his students at the prospect of AI literature. “When one of my graduate students said ‘as an experiment, I’m going to be writing a part of this piece with AI’, the other students became so angry, they wrote letters to me saying how awful this was.”

“There’s a kind of implicit bargain between writer and reader where you know the work that you’re getting is generated by a human being, and I think it felt like an assault on that,” he says. “Reading literary fiction is this incredible Vulcan mind meld with another human being, entering someone else’s consciousness. With AI I’m entering the simulacrum of another person’s consciousness, one degree removed, or many degrees removed. How sad is that by comparison?”

For Hardaker, “I guess it impinges on what we think of as what makes us special, what makes us valuable and unique”. At the same time, the music-generation model she uses “has generated some absolute bangers. I listen to them, unironically, in my car, and I enjoy them quite a lot.”

Could the same happen with literature? Will a machine-authored novel one day take its place among the 100 greatest of all time? Peter Stockwell, professor of literary linguistics at the University of Nottingham, thinks AI may be able to do the basics, but it can’t scale the heights. “If you want something that’s very familiar and very mediocre and entirely functional, it’s amazingly good at that.”

One way to think of language, he says, is as a series of layers, with words at the bottom followed by phrases, clauses, compound sentences, all the way up to narrative structure. “AI is really good at the lower levels. It’s learned lots of our syntactic structures and so everything looks well formed and grammatical. But, the higher up you go, the less good it is.” The arc of a story is particularly hard for AI to get convincingly right.

“If you’ve got an AI to write a narrative, it can do a pretty good job of having a sequence of events and something happen at the end. But it wouldn’t be a very tellable narrative,” he continues. “Nothing startling or interesting would happen. And if there is anything startling, it will generally look like a mistake, rather than a brilliant twist.”

The secret sauce of great writing remains secret – even to the academics who study it. “Linguists don’t understand, really, how language works at its higher levels,” at the level of discourse, storytelling, enchantment. “We can’t build a machine to do something when we don’t know how it works.” We do have some idea of what it might boil down to – and that’s our fundamentally social natures and, tied in with that, the fact that we are “wetware” – human flesh, with its spikes of adrenaline, rushes of dopamine, craving for social contact, all of which find expression in language’s structure and the way we use it.

There are two broad mod

[truncated for AI cost control]