![\"Robot-writing](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2025\/11\/b1a4c9b0427ca48ead4483655e78825d3dc13c8c.jpeg\")
<\/p>\nAs a testament to language change, let\u2019s delve into the tapestry of modern English. By now your alarm bells should be clanging. There\u2019s something off-key about that first sentence, but what? Maybe it\u2019s the wonky metaphor, since delving is reserved for earth or pockets or even the past \u2013 not tapestries.<\/p>\n
Or perhaps the glitch is tone, the prose a little purple. As a hint, let\u2019s meet Dr Jeremy Nguyen, a senior researcher and lecturer at the Swinburne Business School, who last year sifted some 39 million citations from biomedical literature, looking for instances of \u201cdelve\u201d.<\/p>\n
<\/p>\n
Robot-writing has invaded the system \u2013 and notably, AI has its giveaway words. Like, \u201cnotably\u201d.Credit: Getty Images\/iStockphoto<\/p>\n
Yes, an academic delve into delve, a word adored by chatbots, as Nguyen had gleaned, along with \u201ctestament\u201d and \u201ctapestry\u201d \u2013 the AI trifecta when it comes to medical screeds. The survey revealed all three had spiked in usage, with delve appearing in 0.5 per cent of all articles on the PubMed site (a database of clinical writing), compared with a minimal 0.1 per cent in 2022. Tapestry and testament were just as rife.<\/p>\n
Not long ago, examiners relied on piracy software to detect which students were cribbing their analyses of Othello\u2019s fatal flaw, checking whether Student A was vacuuming Wikipedia or Student B was duplicating the work of Student C. No doubt they still do that, except that lately the task has turned messier with robot-writing invading the system, less piracy than conspiracy.<\/p>\n
Nguyen noted delve for its poetic vibe, a Heaney-like glimmer in the AI generative grey. For its aptness too, as the Large Language Models (or LLMs) that chatbot software needs for its vocab are pretty much quarrying our own human material on a Pilbara scale. The minute the Swinburne research blew the whistle, a cheeky post on ChatGPT\u2019s official X account admitted, \u201cDon\u2019t mind me, just doing some delving today.\u201d<\/p>\n
Testament and tapestry also claim a Shakespearean glamour, both gallant bids to counter the more functional nuts and bolts of prose generation. Other prime suspects in the output audit are \u201cvibrant\u201d, \u201clandscape\u201d, \u201crealm\u201d, \u201cembark\u201d, \u201cexcels\u201d, \u201cvital\u201d, \u201ccomprehensive\u201d and \u201cintricate\u201d. Dare you to use all 10 in a sentence.<\/p>\n
Loading<\/p>\n
Or don\u2019t, as you\u2019ll only be accused of being an algorithm\u2019s accomplice, since current examiners are as likely to be scouting for these mechanical Easter eggs, along with the fancier transitions of \u201cmoreover\u201d, \u201carguably\u201d and \u201cnotably\u201d. Not that Nguyen\u2019s paper isolated that last group. Rather the intel came from Google Overview, essentially a robot dobbing on a robot, echoing Roman poet Juvenal\u2019s words of 2000 years ago: \u201cQuis custodiet ipsos custodes? (Who guards the guardians?)\u201d<\/p>\n
Where Juvenal\u2019s satirical dig \u2013 sorry, delve \u2013 was originally aimed at corrupt city officials trusted to root out corruption in 100AD, the question now seems all the more acute in 2025 as a linguistic coup is offering its own surveillance advice. Quis verba verborum factorum custodit? (Who guards the words of the word creators?)<\/p>\n
Mayhem seems imminent. Where\u2019s Philip K. Dick when we need him? As the future unfolds and virtual writers keep rising, I can picture shady students swapping testament for \u201cwill\u201d, delve for \u201cexcavate\u201d, eager to camouflage their imported text, just as jaded tutors will fossick for \u201cnotably\u201d or \u201cembark\u201d because DeepBlue\u2019s Deep Throat told them such words were smoking guns. Moreover, the issue is one intricate tapestry.<\/p>\n