China commentary: Are long articles “AI-written”? Huxiu columnist says reflex to suspect is the real problem

The claim

A columnist for Huxiu (虎嗅) argues that the newest byproduct of AI is not just machine-made text, but a mass habit: people now reflexively suspect any moderately long piece is “AI-written” and ask machines to summarize it. Why read when you can outsource the reading? The author says this reflex corrodes taste and the very ability to judge quality — turning readers into passive consumers who mistake quantity for insight.

What the author is warning about

The piece lays out a blunt distinction. AI can generate competent, roughly “70/100” copy, but you need about “90/100” human judgement to select, edit and elevate that raw material. Without that skill, users will accept AI output wholesale and lose the ability to tell good work from slop — a term the writer uses for low-effort content flooding social feeds. The article cites examples — including a reader who mistook a 2017 sentence pattern for a modern AI trope — to argue that the suspicion itself has become inverted: people now avoid long-form thinking because machines exist to shortcut it.

Broader context and why it matters

It has been reported that large language models are trained on vast web corpora that include human-generated text, which is why many writers feel ambivalent: their words fuel the systems that then displace attention. Reportedly, this debate plays out against a wider backdrop of geopolitical and trade tensions over advanced AI technologies — from export controls on chips to an intensifying US‑China technology rivalry — that shape which models are available where and how they are regulated. The columnist’s final point is cultural: machines may produce lots of readable words, but they do not necessarily produce the pleasure or rigour of good writing. Who decides what counts as taste — and who is willing to read long enough to find out?