The question of whether a person or a machine wrote the text in front of you has moved from novelty to everyday practical concern. The Sydney Morning Herald has published an explainer under the headline "How to spot if something has been written by AI," surfaced this week through Google News' coverage of large language models.
The available source item is a headline-level listing rather than a full text, so the specific tells and techniques the SMH piece recommends are not reproduced here. What the framing signals, though, is the state of the problem: readers are now routinely encountering text whose authorship is genuinely ambiguous, and mainstream news outlets are responding with consumer-facing guidance rather than technical explainers aimed at specialists.
That shift is the story. Detecting machine-written prose is not a solved problem. There is no reliable public tool that labels text as AI-generated with confidence, which is why the advice on offer tends to be human judgment rather than software — reading for the qualities that give a passage away. Guidance aimed at general readers, as opposed to automated detection, is an implicit acknowledgment that the automated route has not delivered.
The stakes run through ordinary life: job applications, student essays, product reviews, dating profiles, news articles, and the messages in your inbox. Each of these carries an assumption about a person being on the other end. When that assumption stops being safe, the cost is not only deception in individual cases but a slow erosion of the default trust that makes reading anything at all worthwhile.
It matters because the ability to tell human writing from machine writing is quietly becoming a basic literacy skill — and right now, no one can hand you a tool that does it for you.