Fact · Voice and AI
In a 2023 study of several widely used GPT detectors, the detectors consistently misclassified non-native English writing samples as AI-generated.
Checked against its source on October 17, 2026
In the source's own words
“Our findings reveal that these detectors consistently misclassify non-native English writing samples as AI-generated”
GPT detectors are biased against non-native English writers, Weixin Liang et al., arXiv (later published in Patterns), 2023, Abstract.
Open the original- What it measures
- The authors compared writing samples from native and non-native English writers across several detectors. Native writing samples were identified accurately.
- What it does not show
- The study tests general GPT detectors on essays. It does not test Turnitin or speech input, so it does not show that dictated text is flagged.