Granta's embarrassment is a case for AI literacy

Note: this whole thing got more embarrassing since I first published it, so I added an update to the bottom.

I’ve been thinking a lot about a guy named Frederick Lorz lately. You might’ve heard of him. He’s the guy who rode in a car for part of the men’s marathon at the 1904 Summer Olympics.

I first learned about him twenty years ago1, but over the last few days I realized that either I heard the story wrong initially or at some point I began to misremember it.

I thought the story of Frederick Lorz was this: He got into a car during a marathon, used it to skip a few miles, then got back out there, kept running and won a gold medal. Because cars were so new in the world, no one could decide whether he cheated or not. He was allowed to keep his gold medal.

The truth, as I learned from reading this Smithsonian article and this article on the Olympics website, is that Lorz rode in the car for 11 miles of the race, was the first across the finish line and was quickly disqualified once people learned the truth. He meanwhile claimed it was a practical joke and he wasn’t even trying to win. The Olympics banned him but apparently later unbanned him.

I’ve been thinking about Lorz because there’s a new situation that reminds me of him. I’m referring to Jamir Nazir and his short story “The Serpent in the Grove,” which was recently published in Granta after winning a 2026 Commonwealth Short Story Prize. Although I should point out that this latest situation reminds me more of my personally recreated version of the 1904 Olympics, rather than the apparent truth of Frederick Lorz and his shenanigans.

I make no secret of my skepticism around artificial intelligence and LLMs in particular. While I believe there are times it can be useful2, I think most uses of AI are bad, including:

• Writing words you want another human to read

• Generating ideas you want to share with another human

• Creating “art.” I put “art” in quotes because I do not think AI can create art.

• “Outlining” or “copyediting” or “brainstorming” or any of the other things people think they use it for.

And so on.

As I said, I am not against all uses for it. I do think AI is the future of search engines (whether we like it or not) and can be useful for certain forms of research, assuming that the person doing the research has the knowledge and wherewithal to know that machines have a tendency to provide false information.

LLMs have the same core issue the internet has had since the beginning: they repeat and generate lies and the responsibility is on the human user to parse the lies from the truths.

If you haven’t heard of “The Serpent in the Grove” or Jamir Nazir or the 2026 Commonwealth Short Story Prize or even Granta, here’s what’s happening:

1. On May 14, the Commonwealth Foundation announced that an author named Jamir Nazir won a literary prize called the 2026 Commonwealth Short Story Prize. He was one of five regional winners, with his story to be published in Granta.

2. Shortly after, people on the internet (mostly on social media platforms like X, Bluesky and Substack) began publicly accusing the story of being partially or entirely created by AI. Controversy ensued.

3. News publications took note, with the controversy being covered in The New York Times, The Atlantic, Lit Hub, The Independent, The Observer, etc. You should be able to find an article about this controversy (scandal? dustup? drama?) from your preferred news organization.

4. Granta has attempted some damage control, releasing a statement3 and adding a disclaimer to the top of the web page for “The Serpent in the Grove.” The Commonwealth Foundation also released a statement on May 19, followed by another statement on May 22. I’ll be discussing the details of these statements below.

As this has been covered in so many places, I considered not piling on. But my intention with Edit History is to talk about how the truth exists on the internet and this is a case where the truth remains unclear. Plus, I have my own thoughts that I want to share.

I’ve read “The Serpent in the Grove” and I am confident it was written at least partially with an LLM. (I’m gonna start referring to it as TSitG so I don’t have to keep writing it out.) However, at no point did I use an AI detector to reach that conclusion. I don’t use those.

As I’ve said in the past, you should be able to detect AI by using your brain. Or you can compare a writer’s past writing to their current writing and look for AI tells that have appeared post-2022.

I consider using AI detectors to be about as lazy and destructive as using ChatGPT to do your copyediting or “writing.” It’s another way of outsourcing cognitive abilities to a machine in exchange for a false degree of confidence.

In the case of Nazir’s story, I believe it to be AI because it’s full of AI writing tells. Most specifically, the story is full of “Not X, but Y.” If you’ve read spent any time reading AI slop (whether willfully or not), you know LLM-generated text almost always uses this writing quirk.

Here are some of the examples of this tell in TSitG:

• “Not the bees’ neat industry or the clean rasp of cutlass on vine, but a belly sound – as if the earth swallows a shout and holds it there.”

• “… not from joy but from endurance.”

• “Not for anything she did wrong, but for how exactly she fitted the life that fenced him in...”

• “…not like a mother, like a judge.”

• “…not to saints or ghosts, but to a living listening.”

These aren’t even all the examples. I could go on, but instead I’ll quote this Futurism article that I thought laid it out well:

All the AI hallmarks are there: negative parallelisms (“it’s not X, it’s Y”), lists of threes, turgid imagery, and nonsensical figurative language that seems a world apart from the precision the judges lauded it for.

So here’s my first point I want to make: you don’t need an AI detector to detect AI. To run a story like this through a tool is to seek needless affirmation (which, again, is my beef with AI. It’s a confirmation bias machine.)

Anyone who is AI literate should read this story and smell an electric rat.

But if you did need another piece of convincing, this is the image the so-called author submitted of himself, that is currently on display on the Commonwealth Foundation website:

As I’ve said before: I know slop when I see it.

Which brings me to my second point: no one involved in this competition is up to the job, if they can’t recognize words or headshots as being AI-generated. It should not have gotten this far because the judges should have read this and said “Yikes, another AI story. Toss that one in the pile.”

I mentioned above that Granta and the Commonwealth Foundation both attempted some level of damage control.

The Commonwealth Foundation’s May 19 statement can be summed up by refuting AI detectors, which I would consider fair if it didn’t also seem to suggest that AI-generated writing can’t be detected without using a detector. Their May 22 statement essentially said that people should stop talking about this because it is hurting feelings:

At the centre of this are real people: striving writers, dedicated judges, and a passionate team. All those who submit to the prize are people with deeply personal stories; writers facing their own complex challenges through the vulnerable tool of writing. Our Prize holds huge value in recognising their merit by bringing these people to the fore. The Commonwealth Short Story Prize is free to enter and anonymously judged, meaning many finalists are in their early literary careers. We hold a duty of care over them, which we take seriously as they contend with global attention that is entirely new to them. We are disappointed to see the work of current and previous writers of the prize called into question. We feel it is vital to recognise the upset felt by all those involved in or affected by the recent discourse surrounding the Short Story Prize. We offer our support to our winners, shortlisted writers, judges, alumni and wider community.

But that’s nowhere near as bad as the Granta statement. Here’s the worst part of what Granta’s publisher Sigrid Rausing had to say:

We showed Claude.ai the story and asked whether it was AI-generated. The response was long, concluding that it was ‘almost certainly not produced unaided by a human’. It ended, however, on a cautious note:

‘The strongest evidence against pure AI authorship is the small number of passages that don’t fit the pattern – particularly the Zoongie morning-greeting and the agricultural paragraph about the acre. Those passages carry the kind of off-shape specificity that models still struggle to produce unprompted. If the story has a human core, it is concentrated there, and the AI has been used to elaborate around it . . .’

This is beyond painful. The official response by the Granta team was to ask an LLM if an LLM wrote the story?

Also, what does this mean? “…almost certainly not produced unaided by a human.” Is that a good thing? Are we saying that it’s bad because a machine and a human collaborated? Or that it’s good because a machine and a human collaborated? What does Granta think it’s saying with this statement? Why the hell did they feed this story to a sycophantic slop machine? Why would anyone admit they asked Claude about this?

I think the answer is simple, and it’s my third point in writing this: if you aren’t literate in AI, you’re going to get scammed by AI.

So far this story has:

• An alleged writer who submitted an AI slop image of himself in place of a headshot

• A group of writers who were unable to identify some of the most telltale signs of AI in a short story

• A lot of people leaning too hard on AI detectors instead of their own eyes and ears and minds

• A high-ranking person in the literary world who thinks it’s a legitimate idea to ask Claude for its point of view

All of this is depressing. But it’s not surprising. And it has convinced me we need AI literacy because otherwise more people are going to get scammed. Deepfakes, fake news, fake accounts. AI scams are already omnipresent. Tell your parents, tell your kids. Don’t get scammed.

People like to say things like “AI won’t take your job, but your competitor using AI will.”

I’d say the bigger truth is: “AI won’t take your money, but a scammer using AI will.”

Now, I’m not saying to go log into ChatGPT and start churning out your own slop. You don’t need to develop a Claude delusion to protect yourself. When I say AI literacy, I mean the ability to recognize something that has been generated using AI.

This general situation has been a concern of mine since before LLMs. People don’t know how to practice the most basic hygiene on the internet. They regurgitate misquotes,4 click ads in Google without knowing they’re clicking an ad, follow scammers on social media, fall for fake news, share phishing links, learn ostensible information from websites whose domain they don’t take the time to inspect, etc.

Is there a solution for this? I don’t know, but here are some helpful resources:

• Wikipedia’s Signs of AI writing

• This guide from the state of California on AI scams

• This Today Show clip starring my buddy Niko Pueringer talking about AI video.

• The Reporter’s Guide to Detecting AI-Generated Content on the Global Investigative Journlism Network

• This MIT guide called AI Detectors Don’t Work. Here’s What to Do Instead.

Am I going as far as to claim that someone scammed Granta and the Commonwealth Foundation? I don’t know. In order to qualify this as a “scam,” I think we would need to better understand some details of this story that are still unknown.

And while I’m reminded of Frederick Lorz, that case was different. Lorz rode in a car, raced across the finish line, was heralded as the winner and then later played it off as a practical joke. But there wasn’t any ambiguity. First he was declared the winner and then he was declared a cheater. He passed himself off as a jokester, which maybe he was.

A foot race is simple. Who crossed the line first? You either won or you didn’t win. And if you cheated, you didn’t win.

Literature is subjective. It’s why certain types of people have no use for it, i.e. Sam Bankman-Fried5 or Andrew Tate6.

Does this prove that AI can write an award-winning story? Or does it remind us that stories winning awards might be a little bit silly?

Regardless of what it means for the literary world as a whole and the idea of prizes and what stories deserve prizes, the biggest takeaway for me is that this entire story is about the need for AI literacy. Don’t get scammed, folks.

And if you do reveal yourself to lack AI literacy, make sure you do so in a slightly less embarrassing way than what happened this time.

UPDATE:

“The Serpent in the Grove” won the big prize. Jamir Nazir did a bizarre interview with The Atlantic during which he claims he did not use AI, but that he does use text-to-speech and that he admires the work of Derek Walcott but cannot name a single poem by Walcott. Meanwhile, his AI slop remains publicly posted (with very little engagement) on his LinkedIn.

Also worth noting that, while he insists he did not use AI, Nazir gives the same tired arguments defending it that you may have heard before:

And on the same topic of that, what about the typewriter? When the typewriter was first invented, writers kicked hell and said, The thing is writing. You’re supposed to use a quill or your fountain pen. And there was a big hullabaloo. We passed that, and then next came new word processors. I can tell you word processors are machine assistants, because they can check the spelling, right? They can search, find and replace, suggest synonyms, et cetera. Modern-day word processing can do almost everything. Where’s the uproar? It settled down. Now I think the same thing will happen with AI.

All very strange. None of my original opinions on this have changed, although “AI awareness” could be a better name for what I think we need, rather than “AI literacy.”