Yes, on average. Published evaluations of commercial detectors have repeatedly found higher false-positive rates on non-native English prose. Second-language writing often has more regular rhythm and a narrower vocabulary — the same features style-based systems read as machine-like.
That makes percentage cutoffs in schools or hiring pipelines especially harmful: they become a stealth language test. Reviewers should open passage-level evidence, compare to prior work from the same person, and never treat a score as proof of authorship.
