Researchers have used top Generative AI models to grade hundreds of undergraduate essays and found that AI only matched human-awarded degree classification around half the time, with AI often failing to accurately assess the best and worst submissions.
This article was originally published on this website.
