Zonder Labs groups defects in workplace AI writing into hidden actors and actions, vague combinations of words, and sentences with an obscured point. Detecting AI authorship is different from judging whether readers can understand a message. A separate editing-tool case study inspired by this work addresses length and clarity before stylistic tells. The comparisons concern the authors’ samples, not universal model performance; they are best used as editing criteria for PR descriptions and work messages.
Back to Tech
AI writing analysis focuses on unclear explanations rather than stylistic tells
Put actors, actions and the reader's needed conclusion ahead of decorative detail.