The article discusses the growing reliance of students on AI tools like ChatGPT for academic assistance, contributing to significant concerns among educators about attention spans and academic honesty. While some teachers adapt by using AI to assist with grading, studies reveal AI's limitations, particularly in accurately assessing student work. Research from the University of Georgia shows that AI grading performed poorly compared to human evaluators, achieving only 33.5% accuracy when grading middle school assignments without a human rubric.
Talk to a teacher lately, and you'll probably get an earful about AI's effects on student attention spans, reading comprehension, and cheating.
...nearly 86 percent of university students use some type of AI in their work.
While AI might save educators some time...the tech isn't even close to cut out for the job, according to researchers at the University of Georgia.
The results were abysmal. Compared to a human grader, the LLM accurately graded student work just 33.5 percent of the time.
Collection
[
|
...
]