Verified65% confidenceBenchmarkExact time
A 2023 study by Stanford University showed that mainstream AI detection tools have a false positive rate as high as 61% for non-native English writers.
3
Sources
65%
Confidence
Long-term
Relevance
7/12/2026
First Seen
Sources
不禁用AI：师生共建课堂契约的教学实验与启示
hackernewshackernews7/3/2026
Related Claims
UnverifiedAI检测工具对非母语英语写作者的误判率尤为突出，学术界已有学生因此被错误指控作弊的真实案例80% similarVerifiedAI detection tools are highly prone to misclassifying texts by non-native speakers and novice academic writers, with false positive rates exceeding 20% in multiple tests.80% similarVerifiedThe false-positive rates of mainstream AI detection tools (such as GPTZero, Originality.ai, and Turnitin) can reach 5%–15%, and the false-positive rate is even higher for academic writing by non-native English speakers.79% similarVerifiedAI文本检测工具对非英语文本的检测能力更弱，且存在将非母语者的英文写作误判为AI生成的公平性问题73% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/489105API
curl https://kongchang.com/api/v1/knowledge/claims/489105MCP
get_claim(id=489105)