UC’s inconsistent AI accusations and repercussions - Extended Version

Source: Mike Mackenzie / www.vpnsrus.com

*Names have been changed.

UC students are facing penalties and potential academic integrity reviews due to AI use in assignments. Some deny using AI altogether, while others employ it without repercussions to complete reports.

A second-year student, David*, who prefers not to be named, expressed shock after receiving a 57.5% grade for an essay, with feedback alleging AI involvement despite his claim of never using AI.

"I have no idea how my tutor reached that conclusion, as there's no evidence that I used AI. I completely refute this accusation," he said.

David, who had suffered a head injury affecting his studies, opted to drop the paper following the incident after his doctors’ suggestions to take only three papers.

He clarified, "I've never used AI for my assignments at UC or elsewhere; it would be pointless, and I wouldn't learn anything."

After withdrawing from the paper, David strongly denied using AI and questioned his tutor and lecturer but found that they weren’t forthcoming about their conclusion of his AI use. He attributed his forgetfulness of citations and sections to his injury, supported by his doctor's advice to take only three papers due to the issues he was having.

"I believe they eventually accepted my explanation because they knew it was true," David said.

Ella*, was also accused of potential AI use in a feedback file for her assessment, when she claimed she had not used any.

Although the grade she received she said she had expected, if she didn’t open the feedback file she wouldn’t have known she’d been accused of AI use.

Ella didn’t counteract the claim, but said the tutor even added a comment saying that one of her sentences was awkwardly written, “which if I did use AI, that shouldn’t be the case.”

“It’s weird. I don’t think they have a consistent process at this point.”

On the other side, Peter* is a fourth-year electrical engineering student who has used AI in many assessments and has never received any ramifications or indications from his lecturers that AI was detected.

“I find it odd that many other people get pinged for it even when they don’t use it and I just get nothing. Makes me wonder what’s really going on,” Peter said.

Peter uses AI to write sections and paragraphs in his reports and adjusts it as necessary. Due to use of AI in quite large sections and believes his lecturers don’t care if they use it.

“Many lecturers [in their faculty] just consider it as a tool to make things easier.”

UC marker, Simon*, explained the process in how they identify AI use without seeing a Turnitin score. As a marker, they cannot see the Turnitin results, but can navigate the use of AI through “suspicious signs”, Simon said.

“These can range from sudden changes in tone, completely different voice in a second essay submission from one submitted earlier in the semester, false (potentially hallucinated) references, or at the worse end fragments of writing referring to 'the authors other publications'.”

But with that said, Simon says they don’t decide marks on these cases “on our own steam”.

“Anything that raises these kinds of red flags is separated out of the pack and referred back to the course co-ordinator who does have access to detection tools. We then follow the direction of the course coordinator.”

Simon said a path is sought forward with the interests of the student in mind, “either by trying to guide them towards techniques and skills that will make for better work next time, or giving them a second chance to demonstrate their own thought.”

The University of Canterbury’s (UC) position on AI supports the use of generative AI to enhance teaching, research, and institutional efficiency. Recognising AI's transformative potential, UC aims to "enrich teaching and learning," "support our world-class research," and "streamline organisational efficiency."

However, UC acknowledges the importance of developing high-quality guidelines to mitigate risks related to academic integrity, research quality, institutional reputation, and equity, ensuring these technologies become "a force for positive change"