The Biggest Tell That Something Was Written By AI

AI-generated writing is increasingly infiltrating everyday communication and professional spaces, from personal texts and work emails to literary magazines and opinion sections, according to an editor in Johannesburg who noticed a distinct, uniform voice in submissions and messages. The author observed that even accomplished writers are adopting AI as a "writing tool" under pressure to produce clean, high-volume content in competitive fields, despite widespread reader distrust. The trend undermines the thinking process essential to writing, as efficiency and frictionlessness replace genuine idea development.

The Biggest Tell That Something Was Written By AI Looks closely and you’ll see that everything is equally off. A few weeks ago, where I live in Johannesburg, a man ran a stop sign and crashed into my Subaru. At the scene he was frantic, unable to gather his thoughts. Half an hour later, I received a lengthy, perfectly grammatical text from him elegantly explaining how he perceived the crash had happened. For a repair quote, I wrote to a mechanic I know, a man who used to text me in curt phrases riddled with shorthand. I got a response using just the same voice as the man who’d crashed into me—the distinctive voice of AI. In surveys, people consistently say they distrust AI-generated writing. But that hasn’t stopped more and more of us from using it in everyday life—to compose work emails and personal texts, to make shopping lists, even to write scripts for arguments with our spouses. “I feel like I’m going nuts,” the writer Jason Koebler complained in the tech outlet 404 Media https://www.404media.co/your-ai-use-is-breaking-my-brain/ , under “the cognitive load” of trying to discern whether every piece of text he reads is real or fake. AI writing is also creeping into our most elite literary spaces— newspapers’ opinion sections https://www.theatlantic.com/culture/2026/03/how-ai-creeping-new-york-times/686528/ , books https://www.nytimes.com/2026/05/19/business/media/future-of-truth-ai-quotes.html , literary magazines https://www.theverge.com/tech/936073/ai-writing-granta-commonwealth-prize . I edit professionally, often working with authors renowned for their prose. Maybe two months ago, I began receiving a kind of submission I’d never gotten before: perfectly clean, without a stray comma; uniform in length, with evenly paced paragraphs and a distinctive tone that was simultaneously breezy and grandiose. At first I was surprised that people who prided themselves on their writing would turn to AI to write for them. Even six months ago, when I occasionally identified a paragraph in a writer’s work that seemed AI-generated, they would apologize. Now some authors tell me they’ve embraced AI as a “writing tool,” no different from spell-check or a laptop. The phrase is protean and euphemistic, covering everything from using ChatGPT to find a quote to having it compose a long essay based on a two-sentence prompt. The reason for the change is simple: Competition in journalism and academia and grant writing and even YouTube influencing is insanely fierce. The edge goes to those who can stand out in a deluge of content, which is achieved through cleanly packaged messaging and sheer volume. Even professional communicators who are confident in their writing and unsure that AI is a perfect replacement are under increasing pressure to use it, so long as they feel they’re doing so within their profession’s boundaries. The Atlantic , for the record, prohibits writers from using AI-generated text unless it’s explicitly identified as such. People who aren’t professional writers are making a similar calculation. AI programs’ efficiency in generating smooth, grammatical text is irresistible, whether you need a savvy sentence in a job application or a line of banter on a dating app. AI-generated writing can easily trick readers, especially if they’re only skimming. Tutorials exist for how to strip the telltale signs of AI use from your writing: Get rid of em dashes, colons, and of course the now-icky “It’s not X; it’s Y” formulations. The problem is that the efficiency and frictionlessness that make AI appealing to writers are the same qualities that make it feel untrustworthy to readers. And readers are right not to trust it. No matter how much we may tell ourselves that AI is just a tool like spell-check, it isn’t. When we use AI to flesh out ideas, we lose the most important part of the writing process: thinking. We tend to believe that efficiency is the highest virtue, the four-hour workweek the ultimate goal. Why sweat over the introductory paragraph of an essay if an AI program can sail over whatever argumentative obstacle you have in the space of 15 seconds? But the effort and the hang-ups are, as they say, a feature of the human thought process, not a bug. When human beings write, we judge ourselves; we stop; we backtrack. In published writing, the traces of this process are erased. But it is the process that makes human writing sensible and meaningful. Many authors describe how, when they’ve finally hit on the right idea, writing feels like going down a water slide; putting one sentence after another becomes easy. When writing is hard, it’s often not just because we are tired, underfed, or inefficient but because our mind is trying to tell us crucial things. How many draft texts to colleagues or family members have we all stared at in frustration, wondering why they don’t feel quite right—until we finally realize that they need to be rethought completely, or not sent at all? When a book I was writing became an almost hopeless grind, I tore up 90 percent of the manuscript; it became a far more honest work for having been halted at a conceptual dead end, forcing me to turn back. AI can’t make that kind of judgment. Even if the companies that design AI programs could make them reason like a human being—a project whose hubris is underrated, given that we don’t fully understand the mechanisms behind our own thought processes—they won’t. After all, users consistently say they want AI models to be agreeable, compliant. That means AIs are reluctant to do what the voices in our head do all the time: interrogate the validity of our premise; misunderstand us, forcing us to explain better; insist that a query is stupid; refuse to answer a question at all. In a study published in March https://www.science.org/doi/10.1126/science.aec8352 , a group of Stanford and Carnegie Mellon University scientists found that top AI models affirm their users’ ideas 49 percent more than humans do in conversation. They also discovered that participants rated more sycophantic answers as “higher quality” and said a sycophantic attitude made them more likely to use AI again. So we end up with canned perfection—writing that can’t really be argued with, because it has no underlying deliberative reasoning process, no train of thought. As I wrote on X recently https://x.com/evefairbanks/status/2049094535603437701 , AI writing is almost impossible to edit, because even when it sounds plausible, a closer look will show that every element is equally off: The tone is bland; individual word choices are baffling; the structure lacks sense; key pieces of the argument are missing; facts are false. Working on AI text, as an editor, is like trying to operate on a body whose skin, muscles, veins, bones, and organs are all compromised. There’s nothing to leave intact, nowhere to begin. Recently I asked ChatGPT Pro—which I paid for to summarize financial spreadsheets; I’m not a hater—whether it was capable of viciously dismantling an argument I’d read about the stock market. “I’m tuned to avoid being casually cruel, contemptuous, or needlessly snide,” it explained. “That doesn’t mean I can’t be sharp, skeptical, funny, or sarcastic … I can absolutely be dismissive of bad arguments, flabby prose, obvious spin, bureaucratic nonsense, or market delusions wearing a little tech-futurist hat.” I wanted to look under the hood a little, to understand the reasoning behind how the program chose its phrases. “Why would market delusions be wearing a ‘little’ tech-futurist hat, rather than a big one?” I queried. The program was incapable of pausing to wonder whether it had chosen the wrong word. Instead, it sought to justify itself in increasingly bewildering ways. “Because ‘little’ makes it funnier,” ChatGPT Pro answered. “The market delusion wants to stride into the room as Prometheus; ‘little hat’ makes it look like a raccoon that found a conference badge.” Why a raccoon with a conference badge? I pressed. Why not a dog, or a frog? “Raccoons live in alleys, storm drains, garbage nights, and morally humid environments,” it wrote. “Frogs have existential agency, while raccoons have logistical agency. A frog at a VC afterparty seems like an omen. A raccoon at a VC afterparty seems like it got on the guest list … Think: alley behind a conference hotel, the loading dock of a casino, the VC afterparty where everyone is saying ‘alignment’ but meaning ‘exit liquidity’ … Alignment sounds like ethics, safety, shared goals, humanity, the future. It has dry-cleaned blazer energy … Exit liquidity is socially clammy.” All of these sentences are grammatically perfect. They also make no sense. And all substantially AI-generated writing is like this, under the hood. ChatGPT Pro agrees, though it has no idea why. “What if I said the whole raccoon thing made no sense as a metaphor?” I asked the program. “I would agree,” it replied, ever servile. “At first, the raccoon worked as a throwaway image … A metaphor can survive one or two explanatory layers. After that, it starts looking like a raccoon in a trench coat.” This is the kind of communication we’re becoming surrounded with. Its infiltration into every domain of our lives can’t be stopped. Even people who don’t use AI will begin sounding more like it https://www.theatlantic.com/technology/archive/2025/04/great-language-flattening/682627/ . A preprint by the Max Planck Institute for Human Development https://arxiv.org/pdf/2409.01754 found that in off-the-cuff verbal conversations, such as podcast discussions, people are already exhibiting “a measurable and abrupt increase in the use of words preferentially generated by ChatGPT—such as delve, comprehend, boast, swift, and meticulous.” After all, we remain so much smarter than machines, so much subtler, and thus so much quicker to learn and pick up cultural cues. The difference in how we operate will be extraordinary, and not at all hypothetical. Ten years ago I composed a reconciliatory email to a boyfriend but never sent it, because I couldn’t get the phrasing right. Only much later did I realize I simply didn’t mean what I’d been trying to write. If I’d had an AI program to help me get over the hump, I’d be married to a different person. A much less suitable one. Maybe human writing will become like cloth-aged cheese or handloom rugs, an artisanal product created effortfully. Maybe we will come to treasure older writing. Herman Melville, George Orwell, Toni Morrison—all authenticated. Writing like this will be a fossil record for a kind of thought process we buried without realizing it. The other night, as I was drifting off to sleep, a 19th-century poem popped into my head: Wynken, Blynken, and Nod one night Sailed off in a wooden shoe,— Sailed on a river of crystal light Into a sea of dew. A children’s rhyme, but it had a new beauty. Or maybe smooth communiqués that arrive on time and betray no confusion, doubt, or internal struggle—that polish up our images as affable, efficient, and universally, if superficially, wise—is what we want. But at least we should know what we’re sacrificing.