AI Writing Style Signals
What it is
AI writing can be detected less by obvious words and more by style signals: stop words, function words, phrase rhythm, perplexity, consistency, and time drift. Human writing is often more variable and less internally consistent than model output.
Why it matters
If style is partly subconscious, a good AI writing workflow cannot rely only on “write in my style”. It needs real samples, style tests, updated examples, and human editing. The goal is not to remove all variance, because variance itself is part of human voice.
Evidence across sources
- Cornell research cited by Every found that stop words and word order are strong authorship signals.
- Bucknell research cited by Every found that a few samples can improve style imitation sharply, but human writing remains more variable.
- Virginia Tech research cited by Every quantified LLM time drift: model language falls behind evolving human vocabulary after release.
Open questions
- Which style signals matter most for Chinese writing and bilingual writing?
- Can Jean’s own writing samples become a reusable style profile without overfitting into a rigid template?
- How should wiki output distinguish “AI draft for structure” from “human voice final draft”?