AI Detectors Flag Human Writing as Machine-Generated, Raising Concerns in Academia

A recent incident at Idaho State University highlights the limitations of AI detection tools. Chemistry undergraduate Lauren Jager had written her PhD application personal statement herself, but every tool she used to check for authenticity flagged it as almost 100 percent machine-generated. To pass the checks, she deliberately made her writing less polished and submitted a statement that she considered inferior to her original. She was later accepted into a PhD program at the University of Utah.

The experience reflects a growing crisis in academic integrity technology. Universities are deploying AI detection tools to police student submissions, but researchers have found these instruments to be unreliable, biased, and easily circumvented. A 2025 study on GPTZero, widely considered one of the most-used detectors, revealed a false-positive rate of around 16 percent on human-written essays.

Another study from 2023 showed that most AI detection tools performed inconsistently on human text and struggled more with output from advanced models like GPT-4 than older systems. Notably, even the US Declaration of Independence has been repeatedly flagged as between 95 and 100 percent machine-generated by these detectors.

Some experts argue that even reasonably accurate detectors should not be used in high-stakes decisions due to the risk of false positives. Mike Perkins, who researches AI’s impact on academia at British University Vietnam, warned that using such tools could lead to unfair outcomes for students. Marzena Karpinska from Simon Fraser University cautioned that while detectors can identify broad trends across large datasets, they cannot reliably determine individual authorship.

A particular concern is bias: a Stanford University study found that AI detectors incorrectly labelled more than half of essays written by non-native English speakers as machine-generated, with an average false-positive rate of 61.3 percent.