Universities Deploy AI Detectors, Students Counter with Evasion Tools
False Positives: Students Penalized Despite Not Using AI
"Inevitable AI: Need for Education on Responsible Use"

On university campuses across the United States, the debate between students and school administrations over assignments involving generative artificial intelligence (AI) is intensifying. Schools are introducing detection tools to identify AI usage, but among students, technologies that can rewrite AI-generated text to appear as if written by a human are emerging, resulting in a battle between offense and defense.

"Transforms Text to Appear Human-Written"... US Universities Locked in Battle Over AI Use View original image


On January 28 (local time), NBC News in the United States reported that universities are deploying "AI detectors" to catch student misconduct, but "evasion AIs" have now appeared that can effectively neutralize these detection functions. These tools, known as "humanizers," are AIs that revise text to make it seem as if it was written by a person. According to NBC, assignments modified in this way are not detected by AI detectors.


In response, AI detection companies such as Turnitin and GPTZero have begun to implement features that track students' browser search histories or writing logs. However, some humanizers are reportedly equipped with functions that type out the revised text directly, in order to evade such tracking.


Amid these developments, controversy is growing over the accuracy of AI detectors. Cases have emerged where students are mistakenly accused of using AI even when they have not. In particular, students from non-English-speaking backgrounds, who tend to use relatively simple sentence structures or formulaic expressions, are reportedly suffering unfair disadvantages.


Aldan Creo, a graduate student from Spain researching AI detection technology at the University of California, San Diego, told NBC, "If you write properly, you end up being accused of using AI," adding, "To avoid AI detectors, I deliberately leave spelling mistakes or sentence errors in my assignments before submitting them."


Brittany Ka, who was enrolled in an online course at Liberty University in Virginia, said she left the school after receiving a failing grade on an assignment for allegedly using AI. She submitted handwritten notes, email screenshots, and messages exchanged with her professor to prove she had not used AI, but her grade was not restored.


As such cases continue, there are growing calls to reconsider the use of AI detectors at universities. Last year, at the University at Buffalo in New York, an online petition demanding the suspension of AI detector use was launched, gathering signatures from about 1,500 students.


Some argue that AI use should not simply be treated as something to be policed, but that universities should accept it as a reality and provide clear guidelines and ethical standards for its proper use.



Eric Wang, Vice President of Research at QuillBot, a company that develops humanizers, suggested, "Instead of just focusing on deducting points, educators should teach students how to use AI, and how to leverage technology without losing their humanity and creativity."


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing