A particular sentence structure has become so closely associated with artificial intelligence-generated content that language researchers and editors are now treating it as a near-definitive red flag for synthetic writing. The construction in question follows a familiar pattern: "It's not just this — it's that," a rhetorical device that AI systems appear to favor with striking regularity.
According to a report from TechCrunch, the phrase pattern has grown so prevalent in AI-generated text that its presence in a piece of writing has shifted from being a mere hint of machine authorship to something approaching a guarantee. What was once considered a stylish rhetorical flourish in human writing has effectively been co-opted and overused by large language models to the point of becoming a telltale signature.
The phenomenon reflects a broader challenge facing readers, editors, and publishers in the age of generative AI. As tools like ChatGPT and other large language models become increasingly integrated into content creation workflows, distinguishing human-written prose from machine-generated text has become an ongoing and evolving challenge across industries.
Language models are trained on vast datasets of human-written text, and in doing so they tend to identify and replicate patterns that appear effective or authoritative. The result is a kind of stylistic flattening, where certain constructions get repeated so frequently across AI outputs that they begin to feel robotic rather than rhetorical.
For journalists, editors, and content professionals, the identification of such linguistic fingerprints has taken on practical importance. Publications and academic institutions alike have sought reliable methods for detecting AI-generated writing, and while dedicated detection tools exist, many remain imperfect. Recognizing recurring syntactic patterns has emerged as one of the more accessible and human-readable ways to spot machine-authored content.
The irony, of course, is that awareness of such patterns may prompt AI developers to adjust how their models write, shifting the game of detection once again. For now, however, the next time a sentence promises that something is not just one thing but another, readers may have good reason to wonder whether a human was actually behind the keyboard.
