The spread of 'AI cheating' is destroying education at an alarming rate



Regarding AI and education, while there is a consistent attitude that cheating by leaving academic work entirely to AI

is not tolerated , it has been pointed out that schools are not keeping up with technological advances, as research has shown that it is difficult to accurately detect cheating using AI. Through interviews with young people who have actually used AI to cheat and educators who are struggling to deal with such students, the overseas media Intelligencer has compiled the current state of higher education, which is rapidly collapsing due to AI.

Rampant AI Cheating Is Ruining Education Alarmingly Fast
https://nymag.com/intelligencer/article/openai-chatgpt-ai-cheating-education-college-students-school.html

The first young person to speak to Intelligencer is Chungin Lee, a former Columbia University student. Born in South Korea and raised in the suburbs of Atlanta, where his parents run a college admissions consulting firm, Lee was accepted to Harvard University while in high school, but was suspended after it was discovered he had secretly skipped a field trip before graduation, which resulted in his acceptance being rescinded.

After failing to get into any of the 26 universities he applied to, Lee spent the next year at a community college before transferring to Columbia University as a sophomore computer science major.

However, in the introductory programming class, Lee relied on AI and submitted almost exactly what ChatGPT output for his assignments. In addition, 80% of the essay assignments were written by AI, and Lee only made changes to the final 20% of the assignments.

When Intelligencer asked Lee, 'Why did you decide to leave your studies at university to AI?' he replied, 'Because university is the perfect place to meet your startup co-founder and future wife.'

In addition, Mr. Lee actually founded an AI startup with a co-founder. You can find out more about that in the article below.

A former college student who was expelled for using AI in job hunting starts an AI cheat tool company 'Cluely' and raises 740 million yen in funding - GIGAZINE



The second is Sarah, a first-year student at Wilfrid Laurier University in Canada. Note that while Lee is a real name, all current students mentioned in this article, including Sarah, are using pseudonyms.

Sarah first tried ChatGPT in the spring of her final year of high school, and after getting used to the generative AI, she started using ChatGPT in all her classes. Thanks to that, her grades improved, and she said, 'AI has changed my life.'

Sarah continued to use ChatGPT after entering university, and claims that ChatGPT allows her to write an essay in 2 hours that would normally take her 12. She also claims that hardly a day goes by without seeing other students using ChatGPT on their laptops during class.

Some students are not happy about receiving an education with such an attitude. Wendy, a finance major at a prestigious local university, is against the use of AI. 'I'm against copy and paste, and I'm against cheating and plagiarism,' she said.

However, even Wendy is using AI to complete her university assignments. Wendy, who is not good at structuring sentences, first inputs 'I am a freshman in college and I am taking an English class' into the AI, then gives background information about the class and prompts the professor's instructions. Then, when she asks the AI to 'make an outline so that I can write an essay according to the prompts,' the AI outputs the outline, introduction, and structure of the essay, and she writes the essay based on that.



As AI becomes more prevalent among students, teachers are trying to find ways to respond, by reverting assignments to handwritten exams, switching to oral exams, or making them subject to the use of AI.

Brian Patrick Green, a technology ethicist at Santa Clara University in the United States, stopped giving essay assignments immediately after first touching ChatGPT. He then gave a book report assignment, thinking, 'No one will use ChatGPT like this,' but one of his students soon submitted a book report with robotic language and awkward phrasing. Another teacher, a philosophy professor at the University of Arkansas at Little Rock, said he found students using AI even for questions like, 'Please give a brief introduction of yourself and what you want to learn in class.'

Similar fraud was reaching a certain peak even before OpenAI released ChatGPT in November 2022. At the time when the COVID-19 pandemic was raging, many college students were graduating from high school through distance learning, and since there was no supervision, many students used paid cheating services such as Chegg and Course Hero. Such cheating tools boasted that experts in India and other countries were available 24 hours a day, 365 days a year and could answer assignments in as little as 30 minutes, but when ChatGPT was released, students looking for faster and more functional tools quickly jumped on it.

This rapid spread of generative AI has left educational institutions overwhelmed, with most universities adopting ad-hoc approaches or leaving it up to professors to decide whether or not to allow students to use AI.

Some educators are using ingenious methods to combat AI-assisted cheating, such as 'Trojan horse' tactics, in which unrelated instructions are slipped into the text of an assignment in small, white text between the lines.

'Sometimes these ideas work,' said Troy Jollimore, a philosophy professor at California State University, Chico. 'For example, I've written, 'How would Aristoles answer this?' for assignments that had nothing to do with Aristotle. Other times I've written even more outlandish things that students didn't notice. This means that not only are they not writing the essays themselves, they're not even reading the essays they're turning in.'

The fact that similar AI countermeasures are being implemented in Japanese educational institutions was once discussed on X (formerly Twitter).



In addition, highly accurate AI detection tools such as Turnitin have appeared, but students are fighting back by deliberately making spelling mistakes, revising parts of sentences themselves, or instructing the AI to 'write as if I were a slightly silly college freshman.' The battle between educators who are trying to detect the use of AI and students who are using it secretly is becoming a cat-and-mouse game.

Mark, a mathematics student at the University of Chicago, told a friend that he had used AI for a coding assignment that he had done. He was told, 'If you were a carpenter and you used power tools to build a house, the house still wouldn't be completed without you.'

'Still, it's really difficult to judge,' Mark told Intelligencer. 'Can I really call it my job to write something that was written using AI?'

Related Posts:

in Software, Posted by log1l_ks