Hidden figures
In a devious but clever experiment, researchers at a UK university submitted AI-generated answers to real tests from their own institution, misleading fellow professors and getting better grades than real students.
The research, conducted at the University of Reading and published in the journal PLOS ONEillustrates how AI is upending education in general, creating new headaches for instructors as they figure out how to educate and assess students in the future.
“Our study highlights the responsibility we have as producers and consumers of information,” psychology professor and co-author of the study Etienne Roesch said in a statement about the research. “We must redouble our commitment to academic and research integrity.”
For the study, the researchers administered exams from several undergraduate psychology courses, asked OpenAI’s ChatGPT to generate answers for them, and submitted the resulting work using fake student names. The exams required either short written answers of up to 200 words or essays of 1,500 words.
Other professors, unaware of the research, graded the exams and cited AI-generated answers in only a mere six percent of cases, according to the article, with the vast majority considered human. And remarkably, the AI answers received higher grades compared to real students at Reading.
What’s next
The researchers call it “the largest and most robust blind study of its kind” to date on the impact of AI on education. They claim the research shows that ChatGPT can pass the Turing test, a method of seeing whether a computer can communicate with a human and trick that person into thinking it is another person.
The researchers in the study also propose changing the way students are assessed. They propose integrating AI into education in some way.
When it comes to assessing students and preventing AI-assisted plagiarism, teachers have already explored their options, from oral exams to classroom work. Take-home exams or even tests done on a computer could quickly become a thing of the past if they aren’t already on the go.
But when it comes to using AI in the classroom, it’s a controversial topic that’s sure to divide educators. They either panic because they think that future students will become helpless and rely solely on AI, or they are optimistic that AI is a cool technological tool with no significant drawbacks.
Time will tell which direction society is heading.
More about AI: In new experiment, young children destroy AI on basic tasks
