A Guardian columnist recently did something a lot of people in creative work have probably thought about, but would rather not test in public: he challenged ChatGPT to a writing competition. What made the piece land wasn’t that the AI produced some breathtaking work. It was that the human writer, Rhik Samadder, came away admitting that by the standards many people now use online, the machine might already be good enough to win.
That is the part worth paying attention to.
This is not really a story about whether AI can outwrite a skilled human being in any meaningful artistic sense. Samadder himself makes clear that he still prefers his own work and sees the AI’s output as shallow. The more unsettling point is that he also believes many readers might choose the AI version anyway, or at least not care enough about the difference. He ends his experiment with a line that cuts deeper than the usual AI panic: his job may be over, but writing itself is not.
That distinction matters. A job is a market transaction. Writing is a human act. Those two things overlap, but they are not the same thing.
The Real Issue Isn’t Better Writing
For years, people assumed automation would first take the repetitive, manual, industrial work. Now we are watching software move straight into fields that many people once believed were protected by taste, personality, and original thought. Not always because the machine is better, but because it is faster, cheaper, and instantly available. In the online economy, that combination can be enough.
That is the real lesson from Samadder’s experiment. The danger is not that AI suddenly became soulful. It is that so much of the modern internet is built to reward output that is slick, quick, and passable. A lot of digital publishing does not reward depth. It rewards speed, volume, search visibility, and content that reads smoothly enough for someone to spend thirty seconds with it before moving on. Under those conditions, “good enough” becomes a serious competitor to “actually good.” That is bad news for writers, but it should also bother readers.
This Doesn’t Stop With Writers
And this does not stop with columnists.
Think about marketing copy, product descriptions, newsletters, press releases, corporate blog posts, social captions, email campaigns, explainer content, and the endless river of informational material that fills the web every day. Much of that work is already vulnerable because the bar is not brilliant. The bar is competent, readable, and cheap. Once that is the standard, AI does not need to be exceptional. It only needs to be serviceable. That is a very different threshold, and a much easier one to cross.
Newsrooms can see this coming. Nieman Lab has reported that many journalists expect AI use in publishing to move beyond scattered experiments and into more automated workflows in 2026, with humans increasingly reviewing and shaping machine-produced material rather than creating every line from scratch. Reuters has also publicly stated that it uses generative AI as a supplementary tool across reporting, writing, editing, production, and publishing, while emphasizing disclosure and human oversight when AI becomes a game-changer.
That doesn’t mean human writers are finished. It does mean the value proposition is changing.
What Still Has Value
If AI can produce a draft in seconds, then the premium is no longer on merely producing words. It moves to judgment, taste, reporting, lived experience, credibility, structure, voice, and knowing what is worth saying in the first place. In other words, the future may belong less to people who can fill space and more to people who can bring meaning, trust, and a point of view.
There is another uncomfortable layer here. Around the same time Samadder’s piece appeared, the Guardian joined a coalition of major publishers pushing back on unpaid AI use of journalism, arguing that original reporting is being absorbed by AI systems in ways that threaten the economics of the press. That makes this moment even stranger: publishers are experimenting with AI because they feel they have to, while also warning that AI companies are feeding on the very work that keeps journalism alive.
That tension is not going away. And it is bigger than journalism.
The same pattern is playing out across white-collar work. The question is no longer, “Can AI do this perfectly?” In many cases, the real question is, “Can it do this cheaply enough, quickly enough, and convincingly enough that somebody decides perfection no longer matters?” Once you see that, Samadder’s experiment stops being a media curiosity and starts looking like a preview.
Why This Matters to Everyone
I do not think human writing disappears because a chatbot can imitate rhythm and produce clean paragraphs. But I do think we are entering a period where the market may pay for less depth than we would like to admit. That is the problem. Not that AI has become truly creative, but that much of the web has trained us to settle for writing that only sounds finished.
And that is why this story matters to more than writers. It is about work, standards, and the quiet lowering of expectations. When “good enough” wins often enough, it does not just change who gets paid. It changes what all of us get used to reading.
Photo by Kaitlyn Baker; Unsplash
