Are Scottish students using AI to cheat their way to a degree?

More than 600 Scottish students were accused of misusing AI during part of their studies last year – a rise of 121% on 2023 figures.

According to a freedom of information request by BBC Scotland, less than 10 students were kicked off their courses at both Robert Gordon and Glasgow universities in the last year – the first time AI-linked expulsions have been recorded at any Scottish university.

So when does turning to the internet to help find the answer to a question cross the line into cheating?

That’s the big issue facing Scottish universities, as day-to-day reliance on generative artificial intelligence (AI) tools, such as Chat GPT, becomes the norm.

Dylan Walch is an education advisor with the University of Edinburgh’s Students’ Association and said the number of staff and students seeking advice around the use of AI was on the rise.

He says students who are already using AI often come to the association to seek clarity around its use, while a minority have been caught misusing it.

He told BBC Scotland News that low-level cases would prompt a discussion with a lecturer which might lead to a small mark deduction, but in more severe cases the university would interview the student in front of a panel.

“It can be quite intimidating and quite a big procedure and they can lose anywhere from 10 to 50 marks,” he said.

Scotland has a student population of 292,240.

Figures among AI misuse remain low, but academics say it poses a real challenge around keeping the grading process “fair”.

Prof Sian Bayne is leading the research around the use of AI in higher education at the University of Edinburgh and says it is a “complex” debate.

“The most high-profile concern has been around misconduct and cheating in assessments by generating essays and coursework,” she said.

“I think one of my main concerns is there’s increasing research which shows that it enables a cognitive offloading, instead of reading a complex text and analysing it, it’s much easier to put a complex text into generative AI and get that generative AI to do that complex work for them.”

Only two universities in Scotland, Robert Gordon and Abertay, have dedicated software to help detect the misuse of AI, yet misuse rates remain higher at other universities – with Stirling recording the highest level at 262 cases in the last year.

“Some institutions are thinking about investing in generative AI detection software,” Prof Bayne said.

She said lecturers were “pretty good” at spotting work turned in using AI, but warned that would become more difficult as the technology evolved.

She said: “At the moment we are thinking about ways which we can redesign assessment so that we’re using more multi-model methods like images, audio and video or more oral assessments or more in-person exams, as other ways of protecting against that kind of misconduct”.

BBC Scotland News met a group of Edinburgh university students who argue that AI serves them more as a tool than a shortcut when it comes to their learning.

Masters student George Karabassis, 26, uses AI to assist with translation as English isn’t his first language.

He said: “If I don’t understand a certain question I can copy and paste a certain phrase and ask Chat GPT, for example; ‘could you please explain this in simpler terms’ so I can understand it.”

Akrit Ghimire, 18, said many of his friends used AI to help with their studies.

He said: “Some of the best prompts that they use is ‘now explain it to me like I’m a five-year-old or a 14-year-old’.”

Others in the group also use AI, but warned about the accuracy of some of the search results that they received.

Law student Hannah Dong, 20, said she had come across false content in her use of AI.

“Sometimes if you search for cases they don’t really exist and even if you do find cases that do exist then AI gives you lack of context,” she said.

The group told the BBC they wanted better guidelines set out from institutions on AI, and its use, to keep everyone on a level playing field.

In less than three years, generative AI has flipped the conversation on technology and education, but the debate on how best to tackle its use in coursework continues.