AI a threat to university degrees in South Africa

University students’ use of artificial intelligence poses a significant threat to the credibility of their degrees and diplomas, and government incentives are further exacerbating this issue.

This is according to Professor Andy Carolin, an associate professor at the University of Johannesburg’s department of English, who believes these fears resonate with academics across South Africa.

“I think that we’re much closer to a crisis than many of us are willing to acknowledge, and it’s worth pointing out that this is not fear-mongering,” Carolin told Cape Talk.

“I have no doubt students use ChatGPT to submit written work, but what sets the large language model apart from others is that there is just no way to prove it.”

He says this is different to plagiarism, which can be traced back to a specific source and compared to a student’s submission, thus proving wrongdoing.

When asked about the effectiveness of AI detectors, such as Grammarly or Scriber, Carolin said they only provide a surface claim of AI use.

However, he argues this is not the only way AI affects tertiary education, as many students are not even attempting to engage with the prescribed course material.

Because of this apathy brought about by AI in tertiary institutions, Corolin said that many students are graduating with degrees, having never been assessed or taught critical thinking.

Carolin highlighted the danger of this in an op-ed published in the Daily Maverick.

“We risk an ever-increasing number of students who hold certificates that fraudulently certify their mastery of skills and content knowledge that some may have only barely attempted.”

He said this crisis affects individual students, higher education more broadly, and the South African economy.

One way Carolin believes academic integrity can be maintained is by having in-person exam venues — something many tertiary institutions have yet to revert to after the Covid-19 pandemic.

He says the challenges of AI use must be located within the broader higher education ecosystem, which is influenced by government incentives that could indirectly exacerbate the issue.

“The government has a particular subsidy model that rewards universities with high pass rates and students who quickly move through the system,” Carolin said.

“Universities then pass the pressure to academics, with some even creating pass rates as part of the performance management system, which impacts job security and bonuses.”

He argued that many universities have switched to continuous assessments and portfolios, which allow students to achieve passable marks for work prepared weeks in advance.

Carolin says this subsidy model makes the “South African higher education system unhealthily receptive to tools that allow students to pass off work they did not write.”

AI use in engineering degrees

It’s worth noting that using large language models like ChatGPT may pose a greater threat to faculties where course material is centred around the study of language and its use.

Like Carolin, Dean of the Faculty of Engineering, Built Environment and Information Technology (EBIT) at the University of Pretoria (UP), Prof Wynand Steyn, told MyBroadband that AI has been a significant disruptor.

However, UP’s EBIT embraces the tool. “I’m quite honest and open about it. We do incorporate it. It’s not a sin to use generative AI because it’s a normal tool out there,” said Steyn.

“If we tell our students, no, you cannot use it at all blanket full stop, we would be doing a major disservice to them because the industry would expect that knowledge.”

Steyn said that artificial intelligence is not just used as a tool to aid work but is also being incorporated into assignments.

Some assignments, for example, require students to use ChatGPT or other generative AI tools to write a two-page paper about a particular topic.

The part they are ultimately assessed on is critically linking that prompt to the curriculum done in the past four weeks.

“Critique that document and tell us why aspects that are in there are either good or bad or correct, or how it can be improved,” explained Steyn.

However, he stressed that AI is not a complete solution, given the level it is currently at, and requires human understanding.

To this end, guidelines are constantly being developed for both staff and students outlining how AI can be used ethically and where its use should be restricted.

Steyn provides an example to illustrate this. While AI may be entirely acceptable to schedule a hair appointment without human involvement, certain situations require human interaction.

For instance, a parent calling to book an appointment with an oncologist for a child with cancer would expect to speak to a human who can listen and respond with empathy.