How a liberal arts college can embrace AI tools

I’m Joe Qin, I’m personally now a data scientist. I’ve been doing data science, perhaps, since my PhD work 35 years ago. My PhD thesis was on neural networks. It’s machine learning that we’re talking about now. So it’s interesting to really see that something that started 35 years ago – it was non-mature, it was not good enough, but as technologies advance, we get something really disruptive, I think. This is where we are now, for this, basically, generative types of AI. For Lingnan University, we are now really embracing this AI revolution for all our disciplines and also make sure that our liberal arts education – that’s whole-person education – that needs to be broad enough and also free enough for students to choose, so if we had only the typical liberal arts programmes, that’s not free enough. We need to give our students the freedom to choose history or they want to do some creative things in digital technologies and also sustainability, let’s say, we are now building up the science side of liberal arts education so we have both arts and science well integrated.

How much does GenAI help you in your work as president as Lingnan University?

Very good question. I actually write more than half of my speech scripts using AI but by doing that it’s not saying that I just get it from AI. I do a lot of iterations and also I throw away – probably 20 per cent of the time – I throw away the AI-generated content. It’s not useful.

That’s also how I feel about AI, that we need to interact with it. It’s not a one-shot process, it’s an iterative process. And also, we have to get our own writing skill still there because sometimes it just doesn’t generate the content you want. So, typically I would think of AI as another level of assistant. If the assistant is doing the things that I want, I‘ll use it, otherwise I’ll just do it all myself. This is similar to how you interact with other people. There are good points from other people, but you don’t take them all in, right – you have your own. That’s how I do it. 

What training or resources should universities offer to help faculty use AI in their teaching or resource?

For universities, teaching and research, those are two different things. So the resource requirement is also very different. For the teaching side and now, all the AI software tools are offered over the cloud, so we can just subscribe to licences and make it available to all students and faculty and staff. So it’s not as heavy in terms of the investment as some of the early IT round of updates, so I think we can handle it pretty easily. But on the research side, the resource could be very heavy, because if someone wants to develop AI, that’s totally different. So, I think for education, we use AI as tools, so that’s pretty affordable. 

Now, on the training side, it takes some thinking, so what we did was that we made ChatGPT available to everyone on campus, including students and staff and faculty members. That was almost 18 months ago. So the best way to train them is to have them put their hands on, but we will also organise seminars or tutorials offered by external specialists, and also we have our internal IT colleagues to do some tutorials. 

So that’s needed, but very quickly – these tools are easy to use and we see that our faculty and students are very quick on learning and picking it up. 

How have you won over faculty who are sceptical or hesitant about using ChatGPT or similar tools?

I think I was given the job to take my position as president on July 1st last year and maybe it was a honeymoon period or whatsoever, so I mentioned a lot about the AI, the GPT and I did not hear much direct resistance. However, I think, I know this is a question, it’s an issue, by talking to colleagues from other institutions and so on. What I think we did right was just to give this access to everyone so they can feel it, instead of fearing it. It’s important for them to know what it is, if they have never used, just read about it, that’s not enough. I think we may have some questions initially, but very quickly it became, like, bottom-up, so our new innovations came from the very bottom-level academic units, not from the top-level management. So that’s kind of surprising, also pleasant surprise.

Can you share examples of how your faculty has successfully used AI?

I can share a few things there. One is, in Hong Kong every three years, we have a triennial programme exercise, which means some of the old undergraduate programmes will be phased out and the government requires us to do that. At the same time, we can propose new programmes and these programmes are proposed from bottom up, from the academic units and all the new programmes they propose have some elements of AI or data science. Even from social science, they call it social data science.

So it is, I can see that it’s taken in and also the creativity coming from the bottom, from everyone on campus, not from the top management.  

That’s one element. On the other hand, our teaching and learning centre has developed a platform for faculty and teachers to prepare classes. We call it IDEAL – that’s the acronym. So what that platform does is that the teacher can just use the interface to access AI, access generative AI, enter in some keywords and so on. Then it will generate a list of prompts relevant to the course he or she wants to prepare. Then click, select the prompts and then you’ll generate basically a teaching plan for a day, for the next class. So you’ll take one minute what used to take almost one day, so that’s another evidence of – this is coming from, certainly, the bottom up. 

The third thing is we have established the school of data science officially as another university-level academic unit. I had planned to do that two years into my job, but I think somehow I had to do it within the first year because people asked me, “Why don't you do it now?” So we’ve got colleagues who are very supportive and I think it’s really a lot of communication, and also providing some level of support. We cannot just talk – that won't do it – but we were able to offer everyone access to GPT and subsequent versions. That’s very effective. 

What has introducing a School of Data Science brought to a liberal arts college and why should other liberal arts colleges embrace data science?

Lingnan University has really been thinking about liberal arts education as an education model. So we really say liberal education is how we educate people. It’s not necessary that all the disciplines we offer are liberal arts. So some of the leading liberal arts universities in the world like Princeton or, even, you can talk about West Point as a military academy, they are liberal arts education but they don’t teach just liberal arts.

So that’s upfront and then there is level of familiarity with technology and for those with liberal arts background. I shared with my colleagues, I said this round of technology revolution is not going away from us, it’s coming closer to us. It’s much easier to use than some of the other generations of technology. So I think that’s again received well by our colleagues. It is easier, it’s actually revolutionising the intellectual side of our human being. So for people who generate content, generate text, generate design, or video or arts or whatsoever, this is a great tool. This is the time to use these tools to help do that faster and more efficiently. So I think it’s actually a better match to the knowledge creation disciplines – better than, perhaps, some of other things that’s not yet revolutionised well by AI.

How can institutions encourage collaboration among faculty to share ideas and best practices using AI?

I think there is a bigger picture here in Hong Kong. I would say that, for me, with a background in the US for more than 30 years, Hong Kong actually adapts faster in newer technology for higher education. Hong Kong, we are under what’s called UGC, and they really have realised this is coming, they have set up funds for universities to develop new tools based on AI and also collaborate. So there’s this funding category called Funds for Innovation in Technology for Education, FITE. That also encourages multiple universities to work together, so we already share our developments together, with the support from our educational side.

And then, within our institution, we have purposely elevated the data science from part of a department to an independent school. That all makes sure that our expertise in the data science school can be easily shared across the university. So now almost all other existing faculties or schools want to work with this School of Data Science to have joint programmes and so on, so it is also providing the university-wide education for AI literacy. Our generative AI common course is being offered by our School of Data Science, so to have some kind of organisational change would help.