ChatGPT and artificial intelligence forces educators to rethink assessments

Educators are still coming to grips with how ChatGPT will fundamentally reshape education by providing more opportunities to personalise learning - and more temptation for students to cheat on their assessments.

The free program has "read" more information than a human could ever process and can answer most questions with relatively error-free responses in seconds.

If it is asked the same question multiple times, it will create a different response each time.

This generative artificial intelligence has been in development for a while but it has suddenly burst onto the public consciousness with the launch of ChatGPT on November 30.

Associate professor in University of Canberra's school of information technology and systems, Dr Abu Barkat ullah, believes the emerging technology has great potential in education, with the appropriate regulations in place.

"Is it a blessing or is it a curse? I am an optimistic person. I see the benefit," Dr Barkat said.

AI can be used to personalise learning by analysing how students are performing and customise lessons or assessments based on students' abilities.

It could be used to ease the workload of teachers, freeing them up for higher-value tasks, something that teachers unions around the country have been campaigning for amid a dire shortage of educators.

"It is not replacing any human. It is not replacing the need for the teaching team or the educators," he said.

"AI is not matured yet. The teacher, when they're making the lesson plans or something using the AI tool, they need to review it."

Dr Barkat said instead of a student relying on one-on-one teacher attention, they could ask an AI program questions or ask for feedback on their work.

AI can be used in education in a variety of ways, according to ChatGPT itself:

It could also be misused by students who could ask it to do their written assessments at the eleventh hour.

University of Canberra vice-chancellor academic Geoff Crisp said the university was updating its policies to include artificial intelligence.

It will also issue advice and information to academics and students around its use.

"Of course, we are concerned about academic integrity here at the university, as every university would be, because it goes to the quality of our degrees and our graduates," Professor Crisp said.

"The use of artificial intelligence without permission or without the knowledge of the assessor is really just another form of contract cheating or plagiarism if it's not referenced and not used with permission."

The Australian National University's centre for teaching and learning and school of computing have been assessing ChatGPT to give guidance to staff.

"We are concerned about recent advancements in technology-assisted cheating, and we continuously monitor and update our policies and practices accordingly," an ANU spokesman said.

READ MORE:

The COVID pandemic prompted universities to consider how students are assessed in an online environment.

AI is taking that conversation a step further.

"One of the things we want to do is make sure we set meaningful assessment tasks for students and not have assessment just seen as something that's sort of a transactional activity," Professor Crisp said.

"We want students to see assessment as part of the learning process and make sure that they get something out of assessment that's more than just providing a standard response or a standard answer back to an academic."

Students could be asked to do more oral presentations, timed assessments, invigilated exam or laboratory-based activities as course conveners look to stamp out cheating through AI.

But some courses are inherently based on essays and other written responses.

For these, universities and schools will be turning to tools to detect when words are written by AI instead of a human student.

Plagiarism detection service Turnitin has a prototype of a program that can scan text and tell the user how many sentences have been written by AI.

Turnitin regional vice-president Asia Pacific James Thorley said the product will be released some time this year.

"We're confident at this stage that we will be able to detect it as things currently stand and that's due to the fact that AI generated text is more predictable," he said.

"You can't read it to the naked eye. But when you analyse it, it's more predictable than human-generated text at the moment."

Banning AI outright is not the solution because today's students will be using the technologies in their future workplaces. Mr Thorley can see a use for AI in analysing students' writing and providing feedback.

"Tools like ChatGPT potentially have the ability to change the way we think about what we're focusing on with writing in the way that the calculator sort of took away the need to remember your times table," he said.

"The key point is that aiming to stop people using it, I think, is probably a fool's errand at the end of the day. It's about how you how you educate students about."

We've made it a whole lot easier for you to have your say. Our new comment platform requires only one log-in to access articles and to join the discussion on The Canberra Times website. Find out how to register so you can enjoy civil, friendly and engaging discussions. See our moderation policy here.