When readers did not know who wrote what, AI writing came out slightly ahead. A New York Times test showed people two samples side by side, one by a human author and one by AI, and asked which they liked more. After 86,000 people had taken it, the reported result was that 54% of them preferred AI.
The test left aside whether people support AI or believe human creativity can be replaced. Because readers judged only the writing, even a narrow result is hard to ignore, and it changes the way writers have to think about their value.
It also puts pressure on two beliefs many people hold: that they prefer human writing, and that they can recognise AI writing when they see it. Readers do not always reward human effort they cannot see, because they respond to the page in front of them. If the writing is clear, smooth and easy to read, many may accept it regardless of who or what produced it.
But the test measured only what readers chose. It did not measure originality, lived experience, taste, judgement or emotional weight, so the result should be taken seriously without being stretched too far. Nor does it show that human writers have lost their value.
Being human is not enough as a label, because the value of human writing has to show up inside the work itself. That could be a story only the writer could tell, or a point of view shaped by real experience. It could be judgement that comes from having seen the problem up close.
The pressure falls hardest on ordinary, polished, replaceable work that lacks them. Clean sentences are easier to imitate, familiar structures are easier to reproduce, and generic insight is easier to flatten into something that sounds acceptable. For people who write for a living, clear writing is the starting point, and the skill that counts is knowing what only they can bring to the page.
Their defence has to become stronger. It cannot rely on the reader caring who wrote a piece before the writing gives the reader a reason to care.
