Top universities snap into action over AI cheating

Major Australian universities are updating their academic integrity policies and examination procedures in response to AI chatbots like ChatGPT.

This image is partially generated by artificial intelligence.

Major Australian universities are updating their academic integrity policies and examination procedures in response to AI chatbots like ChatGPT.

Australia's top universities are updating integrity policies and redesigning exams to account for the risk students will use sophisticated artificial intelligence to cheat.

It comes amid the release of a powerful and publicly-available artificial intelligence chatbot ChatGPT at the end of last month, which can devise unique essays with a single line of prompt text.

Group of Eight (Go8), which comprises the University of Sydney, UNSW, Monash, UniMelb, UWA, ANU, the University of Queensland, and the University of Adelaide, is "proactively tackling the emergence of AI" through redesigning assessments and using new targeted detection strategies, Chief Executive Vicki Thomson said. 

I guess GPT-3 is old news, but playing with OpenAI’s new chatbot is mindblowing. https://t.co/so1TuXMQB0

We’re witnessing the death of the college essay in realtime. Here’s the response to a prompt from one of my 200-level history classes at Amherst

Solid A- work in 10 seconds pic.twitter.com/z1KPxiAc1O

— Corry Wang (@corry_wang) December 1, 2022

"Our universities have revised how they will run assessments in 2023," she said. This includes "Live+" exams for offshore and online students, in which a supervisor monitors their computer screen throughout, more in-person supervision for local students, and greater use of pen and paper exams and tests. They will also limit the use of "Record+" exams, in which a supervisor reviews the computer screen after the exam, for units with "low-integrity risks".

Some of these universities are also "specifically addressing AI in their academic integrity policies and/or their student Code of Conduct", meaning if students breach policies they can be marked down or failed. 

The University of Sydney board approved an update to its academic integrity policy in November to include ‘content generated using artificial intelligence’, which will be live in semester 1, 2023. The University of Queensland also mentions AI as one form of contract cheating in its policy to be implemented next year.

A Sydney University spokesperson said: "At the moment, we’ve seen very few cases where AI is suspected, and the work has been of a very low standard that would not achieve a pass mark. However, we know that could change and are preparing accordingly".

Along with the Go8 revisions, USyd is also trialing “Bring Your Own Laptop” exams, where tests are delivered online but with face-to-face supervision.

Tertiary Education Quality and Standards Agency (TEQSA) Integrity Unit Director Dr Helen Gniel said while some institutions are banning AI through their institutional policies, the long-term response to open-source AI tools would need to be “more sophisticated than just banning them”.

"It’s clear that students could be using artificial intelligence to produce work for them. But whether that constitutes cheating, it's not something that's straightforward to say.”

@spatial_hair

This is not f***ing around. We must proceed with caution. #ai #gpt #chatgpt #unity3d #gamedev

♬ original sound - Logan Smith

"There are some disciplines potentially that will take a different path because increasingly, copywriters are using these tools, for example, so students are going to need to learn how to use them and how to use them ethically, during their studies,” Gniel said.

Gniel said the sector was "alive to the threat (AI) could pose if not handled properly" but added, "it's not something I think is insurmountable for our sector to deal with".

She said essays were just one form of assessment available to universities but alluded to some of the tools they had at their disposal to determine if an essay has been written by artificial intelligence or by a student. “You’d expect the document properties to look fundamentally different. The total editing time, for example … Looking at that metadata provides a lot of clues to these kinds of investigations," she said.

TEQSA has been cracking down on academic cheating websites that offer human-written assignments for a price, shutting down more than 150 since July 2021, but said AI was a fundamentally different challenge because it cannot blackmail students.

"One is increasingly predatory commercial service that carries risks for the individual students. And one is a technology enhancement which is almost un-owned, and that the sector is aware of and has to adapt to."

Flinders University pre-emptively updated its academic integrity policy back in 2019 after noting early AI-generated papers were accepted at international conferences.

"The conversation in Senate was that if you could fool conference organisers, we should ensure we were covered in our policies if students used something similar,” then-Chair of Flinders’ Academic Senate, Professor John Roddick, said.

Adelaide business school ICHM and the Institute of Health and Management (IHM) also have academic integrity policies that include AI as a form of contract cheating.