Liberal arts universities in a ChatGPT era – How to adapt?

Unlike some liberal arts university leaders, worried about declining enrolment as students become concerned about a future where many writing tasks are taken on by ChatGPT-like AI writing tools, Qin – who was previously dean of data science at City University of Hong Kong – is embracing generative AI.

Lingnan has purchased the licence for version 3.5 of ChatGPT for the whole university and Qin is intent on making sure it is used responsibly by all students and faculty.

After a period of initial caution earlier this year when Lingnan was among several universities in Hong Kong to restrict the use of ChatGPT, it now embraces the technology.

In August the University of Hong Kong (HKU) led the way among universities in the city by allowing the use of generative AI but with caveats.

However, Qin sees no need to restrict the number of questions individual students put to ChatGPT, which is something HKU has done. By granting students free usage Lingnan could emerge as a testbed for the liberal arts in the generative AI era.

“ChatGPT is going to transform the current way of doing things, not just in education but also in the workplace,” he told University World News. His aim is to integrate digital learning and the use of data science tools in the liberal arts curriculum and in research as well as university governance.

“We want all faculty and students to be AI and data literate,” he said. “We will be training faculty and students to use it and are figuring out new ways to assess assignments and support learning.”

This clearly diverges from the emerging policy in mainland China, where a Draft Law was presented to the National People’s Congress in August on the use of generative AI technologies – ChatGPT is barred on the mainland. The draft stated that in academic settings the use of such AI-assisted technologies could lead to degrees being revoked.

Liberal arts is at a critical juncture

Qin said the implications of embracing generative AI go much further than its use in student assignments. For example, course content will have to be rethought. Liberal arts education, he said, is “at a critical juncture”.

Unlike China, Hong Kong’s universities have more freedom to develop the curriculum and can move much faster in revamping programmes to keep up with changes. They also have the freedom to try out innovative approaches that incorporate new technologies in higher education.

“Universities, including liberal arts universities like Lingnan, will have to move very fast into this digital era,” Qin said.

“We are entering a transition phase where anything you want to create within the liberal arts is being revolutionised. [Whether] writing, creating a piece of art or music using generative AI, you can get a draft or a version in just a few seconds. But we [humans] are still at the centre of things,” Qin said.

Part of the curriculum improvements he envisages is inculcating “a new mentality that we have to innovate continuously. Innovation must be in our system”.

As Qin sees it, in an era of easier content generation, students need to be involved in research innovation activities. Innovation, he says, requires a change in mentality: “Students need to learn that knowledge is dynamic, it is changing.” And students must learn how to find out at a time when everything is changing.

“We are moving from a liberal arts university to add research as a new dimension. So [we] would want to call ourselves a ‘liberal arts research university’,” Qin said.

“If we send our graduates into the workplace where one [graduate] knows how to use ChatGPT while another one still does things the old way, you can see a huge difference. We need our students to be the former, who know how to effectively and efficiently use these AI-driven generative technologies,” Qin said.

With an eye to the future, he says the liberal arts university will need a new school of data science, more data-based courses and further opportunities for students to combine arts and sciences.

“We’re going to establish a school of data science, in addition to our social sciences, arts and business schools. The school of data science will bring in a new discipline and also make it easy for us to do interdisciplinary research and education,” he explained.

Teaching students how to use generative AI

Qin, who holds a PhD from the University of Maryland in the United States, originally trained as an engineer specialising in automation and is an expert in data-driven fault diagnosis in the semiconductor industry.

As many industrial processes have made a successful transition to automation he sees the same happening with generative AI, describing generative AI tools “as automation for human beings”.

He points to the industrial revolution where the steam engine liberated people from manual labour, followed by electrification and then the information revolution using computers that is moving towards the AI revolution, noting that each revolution increased efficiency by an order of magnitude.

“Some colleagues see this AI or ‘big data’ revolution as Industrial Revolution 4.0. But I see it as an intellectual revolution,” he said.

Qin uses ChatGPT himself. He says it makes him think more critically, ensuring that ChatGPT gives him the kind of information he wants. “I think of AI as a tool, an assistant,” he noted. “My hope is that students also become more discriminating and critical,” he added.

“ChatGPT does not give one answer for one question, It depends on who uses it. So, the user does have a contribution to what is being generated. The ideas come from the prompts,” Qin explained. “I don’t think ChatGPT can do something that is ready for a student to turn in, even for a simple assignment.”

He explained how using AI-assisted tools will be taught. “At Lingnan, we will give students the opportunity to do assignments using ChatGPT and we’ll also give them opportunities to do the assignment [without] using ChatGPT, so they have to do it both ways. How each instructor or professor assigns that ratio of use or not use is up to the individual instructor,” he said.

To back that up students would also be asked to summarise their own thinking and conclusions. In that way, the learning process can remain clear and coherent.

“We certainly have to train students to be honest – you can plagiarise without ChatGPT, but you don’t want them to think that’s a proper thing to do,” Qin emphasised.

“If you turn in your assignment using ChatGPT, even if only a paragraph is generated from ChatGPT, but you have not acknowledged it, then maybe there is an issue,” he said. Students must be trained to acknowledge its use.

He believes that within 10 years generative AI tools will be even more powerful. Liberal arts students will have to be prepared to function well in such an environment.