9 in 10 AI-produced essays fooling university professors, study finds


FILE - The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston. (AP Photo/Michael Dwyer, File)
FILE - The OpenAI logo is seen on a mobile phone in front of a computer screen displaying output from ChatGPT, March 21, 2023, in Boston. (AP Photo/Michael Dwyer, File)
Facebook Share IconTwitter Share IconEmail Share Icon

A British university failed to recognize more than 9 in 10 essays produced by artificial intelligence (AI), according to a study released last month.

The University of Reading graded 94% of submissions, which were created in the “most detectable way possible,” as if they were written by humans, the study found.

The recent rise in artificial intelligence systems, such as ChatGPT, poses a fundamental problem for the educational sector. In universities and schools, many forms of assessment, such as coursework, are completed without invigilation,” the researchers wrote. “Therefore, students could hand in work as their own which is in fact completed by AI.”

The grades of the AI-produced essays were on average half a grade higher than those written by humans, according to the study. There was an 83% chance the AI submissions would outperform human ones, the researchers noted.

AI’s applications in schools has sparked concern in the U.S., where lawmakers introduced a bill in December to integrate AI literacy as a key component of digital literacy.

It’s no secret that the use of artificial intelligence has skyrocketed over the past few years, playing a key role in the ways we learn, work, and interact with one another,” Rep. Lisa Blunt Rochester, D-Del. and one of the bill’s sponsors, said. “Like any emerging technology, AI presents us with incredible opportunities along with unique challenges.”

In May, a group of senators released a report on how to drive American innovation in AI, in which it warned of risks the technology posed. The senators argued AI could be an asset to the U.S. if handled responsibly.

“In the midst of rapid AI advancements, the Senate can lead or be led. We plan to lead, to deliver for the American people, helping ensure that AI comes as a benefit to society, not a threat,” Sen. Martin Heinrich, D-N.M., said. “This roadmap positions us to unlock AI innovation that will deliver major scientific and medical advancements and help maintain our global leadership.”

Have questions, concerns or tips? Send them to Ray at rjlewis@sbgtv.com.

Loading ...