They Don’t Matter, and Will Soon Be Obsolete

[Pelican Press 0]

This is the hidden content, please

• Sign In
• or
• Sign Up

They Don’t Matter, and Will Soon Be Obsolete

Talk to a teacher lately, and you’ll probably get an earful about AI’s effects on student attention spans, reading comprehension, and cheating.

As AI becomes ubiquitous in everyday life — thanks to tech companies

This is the hidden content, please

• Sign In
• or
• Sign Up

— it’s probably no shocker that students are using software like ChatGPT at a nearly unprecedented scale. One study by the

This is the hidden content, please

• Sign In
• or
• Sign Up

found that nearly 86 percent of university students use some type of AI in their work.

That’s causing some fed-up teachers to fight fire with fire, using AI chatbots to score their students’ work. As one teacher mused

This is the hidden content, please

• Sign In
• or
• Sign Up

: “You are welcome to use AI. Just let me know. If you do, the AI will also grade you. You don’t write it, I don’t read it.”

Others are embracing AI with a smile, using it to “tailor math problems to each student,” in one example

This is the hidden content, please

• Sign In
• or
• Sign Up

. Some go so far as requiring students to use AI. One professor in Ithaca, NY,

This is the hidden content, please

• Sign In
• or
• Sign Up

ChatGPT’s comments on student essays as well as her own, and asks her students to run their essays through AI on their own.

While AI might save educators some time and precious brainpower — which arguably make up the bulk of the gig — the tech isn’t even close to cut out for the job, according to researchers at the University of Georgia. While we should probably all know it’s a bad idea to grade papers with AI, a

This is the hidden content, please

• Sign In
• or
• Sign Up

by the School of Computing at UG gathered data on just how bad it is.

The research tasked the Large Language Model (LLM) Mixtral with grading written responses to middle school homework. Rather than feeding the LLM a human-created rubric, as is

This is the hidden content, please

• Sign In
• or
• Sign Up

in these studies, the UG team tasked Mixtral with creating its own grading system. The results were abysmal.

Compared to a human grader, the LLM accurately graded student work just 33.5 percent of the time. Even when supplied with a human rubric, the model had an accuracy rate of just over 50 percent.

Though the LLM “graded” quickly, its scores were frequently based on flawed logic inherent to LLMs.

“While LLMs can adapt quickly to scoring tasks, they often resort to shortcuts, bypassing deeper logical reasoning expected in human grading,” wrote the researchers.

“Students could mention a temperature increase, and the large language model interprets that all students understand the particles are moving faster when temperatures rise,” said Xiaoming Zhai, one of the UG researchers. “But based upon the student writing, as a human, we’re not able to infer whether the students know whether the particles will move faster or not.”

Though the UG researchers wrote that “incorporating high-quality analytical rubrics designed to reflect human grading logic can mitigate [the] gap and enhance LLMs’ scoring accuracy,” a boost from 33.5 to 50 percent accuracy is laughable. Remember, this is the technology that’s supposed to bring about a “

This is the hidden content, please

• Sign In
• or
• Sign Up

” — a technology we’ve poured

This is the hidden content, please

• Sign In
• or
• Sign Up

than any in human history.

If there were a 50 percent chance your car would fail catastrophically on the highway, none of us would be driving. So why is it okay for teachers to take the same gamble with students?

It’s just further confirmation that AI is no substitute for a living, breathing teacher, and that isn’t likely to change anytime soon. In fact, there’s mounting evidence that AI’s comprehension abilities are getting worse as time goes on and

This is the hidden content, please

• Sign In
• or
• Sign Up

becomes scarce. Recent reporting

This is the hidden content, please

• Sign In
• or
• Sign Up

found that the latest generation of AI models hallucinate as much as 79 percent of the time — way up from past numbers.

When teachers choose to embrace AI, this is the technology they’re shoving off onto their kids:

This is the hidden content, please

• Sign In
• or
• Sign Up

,

This is the hidden content, please

• Sign In
• or
• Sign Up

, and prone to spewing outright lies. That’s before we even get into the

This is the hidden content, please

• Sign In
• or
• Sign Up

that comes with regular AI use. If this is the answer to the AI cheating crisis, then maybe it’d make more sense to cut out the middle man: close the schools and let the kids go one-on-one with their artificial buddies.

More on AI: People With This Level of Education Use AI the Most at Work

This is the hidden content, please

• Sign In
• or
• Sign Up

#Dont #Matter #Obsolete

This is the hidden content, please

• Sign In
• or
• Sign Up

This is the hidden content, please

• Sign In
• or
• Sign Up