Can Humans Detect AI? Mining Textual Signals of AI-Assisted Writing Under Varying Scrutiny Conditions
For researchers and practitioners concerned with AI detection in writing, this study reveals that human judges can detect subtle cues of AI assistance that feature-based methods miss, but the effect is small and incremental.
This study investigates whether warning writers about AI detection changes their writing style and whether human judges can detect AI-assisted writing. Results show judges selected warned writers' documents as human 54.13% of the time vs. 45.87% for unwarned writers (p=0.000243), despite no measurable differences in text features.
This study asks whether the threat of AI detection changes how people write with AI, and whether other people can tell the difference. In a two-phase controlled experiment, 21 participants wrote opinion pieces on remote work using an AI chatbot. Half were randomly warned that their submission would be scanned by an AI detection tool. The other half received no warning. Both groups had access to the same chatbot. In Phase 2, 251 independent judges evaluated 1,999 paired comparisons, each time choosing which document in the pair was written by a human. Judges were not told that both writers had access to AI. Across all evaluations, judges selected the warned writer's document as human 54.13% of the time versus 45.87% for the unwarned writer. A two-sided binomial test rejects chance guessing at p = 0.000243, and the result holds across both writing stances. Yet on every measurable text feature extracted, including AI overlap scores, lexical diversity, sentence structure, and pronoun usage, the two groups were indistinguishable. The judges are picking up on something that feature-based methods do not capture.