Study Reveals Higher Grades for Students Using AI in Exams
A study on artificial intelligence has shown that students who use AI to answer questions receive higher grades in exams. The study, which revealed that even experienced evaluators might struggle to detect answers generated by AI, has raised questions about the academic integrity of universities and higher education institutions.
Dr. Peter Scarfe and his team from Reading’s School of Psychology and Clinical Language Sciences used ChatGPT-4 to generate answers to exam questions and submitted these on behalf of 33 students. The results, evaluated by experts unaware of the study, showed that in many undergraduate psychology modules, AI usage was not detected in 94% of the cases and, on average, received higher grades than genuine student submissions. Dr. Scarfe stated that the findings should be a wake-up call for educational institutions, noting, “The data in our study shows it is very difficult to detect AI-generated answers.”
Dr. Scarfe, who published the research in the journal PLOS ONE, pointed out that AI performed particularly well in the first and second years, emphasizing that as AI continues to develop and become more sophisticated, the education sector must also adapt and update itself. He highlighted the need for universities to focus on understanding how to embrace the new normal of AI to enhance education, stating, “Our research shows that understanding how AI will affect the integrity of educational assessments is of international importance. We won’t necessarily have to fully return to handwritten exams, but the global education sector will evolve in the face of AI.”
Prof. Etienne Roesch, the co-author of the study, emphasized, “As a sector, we need to decide how students should use AI and to what extent we will accept the role of AI in their work. The same applies to the broader use of AI in other areas of life to prevent a crisis of trust across society.”