In a twist that’s both fascinating and concerning, artificial intelligence is not only acing college exams but also outperforming human students, according to a study from the University of Reading.
Published in the journal PLOS ONE, the study conducted a modern-day “Turing test” by submitting AI-generated exam answers alongside those from real students in five undergraduate psychology modules. The results were astonishing: 94% of the AI-crafted submissions went undetected by examiners, despite being entirely AI-generated without any human modification.
Not only did these AI submissions pass unnoticed, but they also consistently outperformed their human counterparts. On average, AI-generated answers scored half a grade boundary higher than those of real students. In some cases, the AI advantage was nearly a full grade boundary, with AI responses achieving first-class honors while human students lagged behind.
This discovery raises critical questions about the future of education and assessment. As AI tools like ChatGPT become more advanced and accessible, universities face the challenge of ensuring the integrity of their exams and the value of their degrees. The study suggests that current methods of detecting AI-generated content are woefully inadequate, leaving educational institutions vulnerable to a new form of high-tech cheating. […]
– Read More: notthebee.com
