Can we "force" students to use AI?

We need to break this down into at least two parts: course content, and teaching choices. From a course content perspective, your colleague may be correct that it is important to include GenAI in the course so that students are prepared for future professional experiences. This is of course dependent on the subject, and if your students do not generally go into one particular career, it may also be guesswork about how professional life will evolve. If your students are all going to be, say, bridge engineers, computer programmers, or nurses, you may already have some ideas about how specific jobs are using GenAI. However generic chatbots like ChatGPT or Copilot are still unlikely to be the real future of professional applications of Large Language Models (LLMs). I think closed-system products are more likely to be helpful for generating options and suggestions for professionals. You don't need the whole internet, you need, for instance, a set of really good peer-reviewed guidance for treatment options in healthcare, or protocols for working with clients with complex circumstances in social work, or photos and frequent problems with spare parts in bridge construction. So it might be very useful to try generic chat products and then discuss with students how they can tell whether the outputs are good and useful and worth the resource consumption, or not. If the material is considered to be essental for course content then it can be compulsory, as long as it is in the course plan (in Sweden) and probably official documents in other countries too. Course plans take a while to change, so there may be a grey area of time when it isn't really compulsory and then students should be able to opt out, but they certainly need to understand why your colleague thinks it's important enough to want to include it.

Now, your colleague may be making teaching choices in order to make their teaching more relevant to the subject, to encourage discussion about these products, and to give students skills that they think are important. Or maybe they like new shiny technology and it motivates them to use it in class. In this context, it shouldn't be compulsory for students to use the products, but they need an alternative activity of equivalent challenge and critical thinking.

Either way, the key to resolving these dilemmas is to have open conversations! We can agree to disagree. We make choices about our consumption every day and sometimes we compromise. The way we humans make those decisions is the important part of the discussion.

The description of your colleague actually applies to me quite well. I have reflected on whether it is legitimate to “force” students to use generative AI, and while I do not consider this an easy question that I have 'solved', I'll try to share some of my perspectives. 

First, I want to consider the pragmatic perspective. That is, what does use of generative AI allow us to do that we otherwise cannot?

The pragmatic perspective can be considered from a ‘professional’ and ‘educational’ perspective. Professionally, at least in my own discipline (mathematics), there is no doubt of the relevance of generative AI. Not only is generative AI able to solve research level problems, but it has also been used to (partially) solve one of the most famous and long standing problems in all of mathematics - The Navier-Stokes problem. Educationally, it seems clear to me that generative AI is a double-edged sword. By this I mean that productive use can supercharge learning (e.g., by providing personalised explanations), but it can also undermine the learning process (e.g., by such personalised explanations which risk removing the ‘productive struggle’ necessary for learning).

Given that many students are likely to use generative AI as part of their studies (whether we like it or not), and are highly likely to encounter generative AI when they enter the workforce, is it ethical of us, as teachers, to leave them to separate productive from unproductive use on their own? 

Pushing the ethical dimension further, one could also argue that we need to help students engage critically and responsibly with generative AI because the technology is so controversial.

I see this latter point as being aligned with my identity as researcher and teacher at the university level. My job is to systematically and critically approach difficult problems (ideally) of relevance to society, where knowledge is both the tool and goal. While personal beliefs, emotions, and prejudices should not be completely dismissed, they should be constantly challenged through critical and responsible engagement — just as we must be willing to challenge our ‘a priori’ knowledge (and methodology). To me, being a researcher is kind of like being Indiana Jones, with the thrill of exploring the unknown coming at the cost of intellectual, emotional, and physical discomfort. And as teachers, our jobs is to invite students to take part of this process (appropriately scaffolded, of course).

But does this I mean that I think we should “force” students to use generative AI? In my class, I try to sidestep this issue by giving the students free choice of which AI to use (and what to use it for). I explicitly state that students are free to choose small AI models, which they can run locally on their own computer (more than 2 million different models are currently available). When the environmental cost is brought down this way, I perceive the issue to be similar to whether we should allow students to refuse other uses of technology in our courses. In this case, I think it comes down to the specific learning goals. If skilled use in the specific technology is a learning goal, then they should not be allowed to refuse. But if the actual use of the technology itself is less important (than, say, manual mastery of related concepts), then refusal may be OK. 

To conclude, I think that exactly because AI is controversial, we need to engage, not stick our heads in the sand. But we should do this for good reasons. If students push back, enter into a dialogue with them and try to find ways to give them room to manoeuvre. And work to make the point of why AI is being used visible.

I think you bring up a very relevant and interesting issue which I don't think we as teachers or the institutions we teach at currently consider enough - the sustainability impacts of using genAI. These issues are also relevant for other aspects of modern living which is very much reliant on digital infrastructures and systems. These require rare metals and energy to create and run, are often designed for obsolescence so they have to be replaced within certain time spans, and can have negative impact to environment as well as to our society. But I'm getting a bit off track here...

When it comes to 'forcing' students to use generative AI (or any digital tool for that matter), I think it is important that the question should asked: what is the underlying learning outcome in this situation or simply, what do I want the students to learn? In my experience from teaching courses within structural engineering I've found it difficult to answer that question with 'to learn how to use a certain tool'. It may, of course, be that this is not the case in other fields. However, I dare say that it is common within many disciplines that the specific tools that professionals use likely change over time - so a necessary learning outcome may be to learn to adapt to this change and get a deeper understanding for what it is that these tools actually help with doing and, more importantly, where the human element factors into this equation (i.e., why do we need the human operator if we are only interested in the output from the tool?). As a researcher, I've previously delved into this issue (pre-genAI) in relation to my discipline. Over multiple decades, the paradigm within professional structural engineering practice has shifted towards a greater reliance on advance computational software. This software, as with genAI, was (and is) becoming more sophisticated and user friendly, meaning that the threshold for being able to use them (in terms of disciplinary knowledge, judgment and competences) was decreasing over time (while the complexity of what they could help create increased). As a teacher, I noticed that students using these tools would often ask questions like, 'how do I get the program to work', rather than, 'is this a good solution'. My approach was to put more emphasis on improving students conceptual understanding of structural behaviour and analysis, giving them the opportunity to develop, appreciate and apply sound engineering judgment, as well as generally adopt a holistic, critical and inquisitive approach. This provided them with the knowledge, competence and judgment to be able to evaluate whether the output from these software was good enough, if there were any errors (e.g., due to poor assumptions, boundary conditions, or input errors), and what could be done to improve the overall design. I've often wondered whether there is a parallel with generative AI - i.e., something that allows students to be able to effectively and with confidence assess the output of generative AI to know when it is good or not, whether it was the right question(s) to ask, knowing what to do with the result(s), etc?

In any case, I don't think we should ban genAI (as this is a bit naive) and I am sceptical to 'forcing' students to use it (at least in my field). If, however, I were to 'force' this onto the student, I would need to be able to clearly and confidently communicate with the students why this is the case (and be myself convinced of this) and also think carefully if there are any other ways to fulfil the expected learning outcomes in the course? As with a lot of other issues in teaching and learning, you can go a far way by talking with your students.

Provided that you have a good reason for students to engage with GenAI, and depending on which part of engagement with GenAI your student is objecting to (and I am assuming that it is about actually prompting a model and thereby contributing to the environmental costs that are triggered by that) and what the indended learning outcomes are, there might also be work-arounds like

• working in groups so that the student who does not want to prompt themself is still discussing prompts and outputs with their peers, but someone else is the one to actually "press the trigger" (but if I were the student who does not want to prompt, this would not satisfy me)
• working with prompts and corresponding outputs that you provide and that the students still analyse critically, or fact-check, or whatever you wanted them to do with it (and in that case they are working with a source you provide like they would with any source you provide)
• critically discussing a prompt that students write and what kind of answer to expect without actually submitting it to GenAI

In the end, I think the best way forward is to discuss their concerns with students and find ways with them to learn what they need to learn while respecting the boundary they articulate.