|
There’s a strange little art to spotting AI writing. Large language models enjoy, for instance, using odd phrases like “strange little art”. Yet some of AI’s writing tricks are trickier to catch. Many people believe LLMs stuff their prose with unnecessarily long, grandiloquent words—and use too many em-dashes. Our analysis of ChatGPT, Claude, Gemini and Grok, looking at 55,940 sentences and 1.2m words, reveals how bots actually write. To learn their peculiarities, and how to avoid sounding like a machine,
read our study.
| 
