AI-Generated Exam Answers Largely Undetected, Outperform Students

Study reveals 94% of ChatGPT-generated responses went unnoticed and scored higher than student submissions, prompting calls for educational reform.

University of Reading conducted the largest study on AI in exams, using ChatGPT for psychology assessments.
AI-generated answers were undetected in 94% of cases and often scored better than real student work.
Researchers urge global education sectors to develop new policies addressing AI's impact on academic integrity.
Current assessment methods may need to evolve to incorporate AI and ensure fair evaluation.
The study highlights the necessity for academic and research integrity in the age of AI.