There is a new step in writing that nobody taught us. You finish a paragraph and read it again. Then you start taking things out. Too many dashes. The one sentence you were pleased with, the one that turns neatly at the end, is now the sentence that makes you nervous. What you end up with is not a better paragraph. It is a safer one.
The writing itself has not gotten any shorter. What has been added is a second stage, where you go back over the whole thing to make sure it does not look like AI wrote it. The em dash is the most famous victim. Writers have used it for 200 years. Over the last year it has become something closer to a confession. Colons are under watch too, and now the pair of commas you use to tuck a clause into the middle of a sentence. People keep lists of these giveaways and pass them around.
Punctuation is a small part of it. The bigger loss is that we are learning to write badly on purpose. AI detectors are built around the idea that machine writing has predictable patterns. So does careful human writing. A controlled vocabulary and a sentence that lands where you aimed it are the marks of someone who edited well. They are also the marks a detector counts against you. Writers respond by putting the mistakes back in. Leave the typo alone. Pick the weaker word, because the better one now turns up in so much published writing that it no longer feels safe to use them.
If English is your second language, you pay twice. The first bill comes from the detectors. A Stanford study ran seven popular detection tools over TOEFL essays. More than 60 per cent were flagged as AI. Every essay had been written by a human being. The same tools were almost never wrong about native English writers. Writing in a second language tends to be plainer and steadier, which is what these tools read as artificial. The exam answer we were taught to write, with its tidy opening and its clean conclusion, is close to the shape that a detector distrusts. The second bill comes from the tools themselves. Cornell University researchers ran an experiment last year with writers from India and the United States. When the Indian participants wrote with AI suggestions, their sentence structure and word choices moved towards American style. We are squeezed from both sides. The machine pulls our writing towards someone else’s norm, and the suspicion punishes us when we do not get there on our own.
The suspicion has reached book publishing too. In May, the detection tool Pangram was run over the five regional winners of the Commonwealth Short Story Prize and flagged three, including Mehendi Nights by the Indian winner Sharon Aruparayil, which it scored as 88 per cent machine written. The organisers asked the writers for working drafts and time-stamped files, and concluded no AI had been used. In August, Nigerian author Jerry Falade’s crime novel was withdrawn from sale by his own agents. Falade denies using AI and says the scrutiny was racial. A detector says a text resembles machine writing, which is not the same as saying who wrote it. The burden lands on the author to prove something that no tool can prove either way.
We do not need a study to see where this ends. Ten newsletters and 10 posts on a feed, and they share one cadence. Some of that is the machine writing. Some of it is people writing to prove they are not the machine. A few now describe themselves as writers from before ChatGPT, as though the year you learnt the craft is a credential. Two opposite fears, and you cannot tell them apart on the page.
Writing is how we find out what we believe, and an instrument you are afraid to use properly stops being an instrument. A whole generation is being taught to read its own fluency as evidence against itself. A few universities have turned them off, having decided they are too unreliable to serve as proof. The detectors can be switched off. The habit will not be. Stop doing the second job. Write the sentence you meant. Keep the em dash if you want to. Let the suspicion be somebody else’s problem.
Hasan is a lawyer
Media365
