A Guardian feature explores how large language models are changing both machine and human writing, asking whether readers can reliably distinguish AI-generated prose from human work. Linguists and novelists describe tell-tale patterns in LLM output, the spread of “AI-sounding” words like “delve”, and growing paranoia in publishing as accusations of undisclosed AI use mount.
This article aggregates reporting from 2 news sources. The TL;DR is AI-generated from original reporting. Race to AGI's analysis provides editorial context on implications for AGI development.
This piece isn’t about a new model; it’s about the cultural wake those models are leaving in their path. As LLM prose saturates everything from hotel reviews to book drafts, the line between human and machine style is blurring to the point that even experts – and detectors – struggle to tell them apart. That erosion of linguistic “tells” means that public trust in text as an indicator of human intent is under real pressure, well before systems exhibit anything like general intelligence.
For AGI watchers, the significance is twofold. First, the feedback loop between humans and models is tightening: people pick up model quirks, models are retrained on human-modified AI text, and the ecosystem converges on a shared, synthetic style. That could make future systems both more persuasive and harder to audit. Second, the arms race around authorship verification in literature foreshadows similar fights over AI-generated scientific papers, legal filings and policy documents. How societies choose to value, certify or penalise AI involvement in those domains will heavily influence how aggressively AGI-scale systems are deployed into high-stakes decision making.



