"AI’s final grade is 99.36, which is higher than both the cohort’s mean (97.70) and median (98.53). "
In the US, graduate student loans have no cap, whereas undergraduate loans have a maximum some where near 55K. I am not sure this study highlights the potential for student cheating using ChatGPT or, instead, the potential for massive abuse of graduate school financial aid. These passing students can borrow towards their housing(mortgage included),food, car and all kinds of tangible property, do not seem to be doing challenging work and seem to have zero chance of not moving on to the next course. As it is healthcare related, they may even qualify for loan forgiveness. I taught for quite some time...the grades and pass rate of this course should merit a thorough investigation outside of the use of ChatGPT. Maybe DOGE needs their first assignment?
the Health Admin course study echos the results of the A real-world test of artificial intelligence infiltration of a university examinations system: A “Turing Test” case study performed at the University of Reading (UK); my takeaway also was that IF writing is used, more than human detection is needed to level the playing field for students (though in that case the mean/median weren't A+ already) https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0305354
"AI’s final grade is 99.36, which is higher than both the cohort’s mean (97.70) and median (98.53). "
In the US, graduate student loans have no cap, whereas undergraduate loans have a maximum some where near 55K. I am not sure this study highlights the potential for student cheating using ChatGPT or, instead, the potential for massive abuse of graduate school financial aid. These passing students can borrow towards their housing(mortgage included),food, car and all kinds of tangible property, do not seem to be doing challenging work and seem to have zero chance of not moving on to the next course. As it is healthcare related, they may even qualify for loan forgiveness. I taught for quite some time...the grades and pass rate of this course should merit a thorough investigation outside of the use of ChatGPT. Maybe DOGE needs their first assignment?
the Health Admin course study echos the results of the A real-world test of artificial intelligence infiltration of a university examinations system: A “Turing Test” case study performed at the University of Reading (UK); my takeaway also was that IF writing is used, more than human detection is needed to level the playing field for students (though in that case the mean/median weren't A+ already) https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0305354