The Problem

University teachers today face a two-layered uncertainty [@KrV_e]:

[Is the old slow and hard way to acquire core virtues and capacities still the way to go, or can we flatten the path of students and allow them to offload the cognitively heavy lifting to AI?]{.underline} For instance, when teaching a philosophy class, should you require (or implore) students to read the complex, and for beginners frustrating, primary texts that make up the most important works of Western philosophy by themselves, or encourage usage of GenAI to get efficient, digestible, tailored summaries? We know a way that works to teach core philosophical virtues and competencies, tolerance of ambiguity, the ability to cognitively penetrate complex conceptual fields, the capacity to think coherently and soundly over long chains: Let them struggle with complex texts, try to write and talk about them, again and again, give them feedback, coach and encourage them along the way, and gradually these virtues and abilities will form. But maybe AI provides a convenient shortcut?
Are these core virtues and abilities still needed in the first place, or will graduates in the workspace of the future need very different ones? While the first uncertainty concerns the path to reach a certain goal, the second one concerns the goal itself. Now, currently, the indicators point in a rather clear direction: Whether it is in software engineering, legal practice, or marketing, what has so far changed is not the ultimate competence profile of senior persons in the field; what has changed is (1) that companies have cut junior roles that had allowed graduates to finish acquiring these virtues and conpetencies because they see AI as a cheaper and comparably able resource to complete these tasks, and (2) that certain services, such as translation, have become much more affordable due to AI. Still, could this not change in the future?

The two uncertainties are of different kinds. The second one is a worry about the emergence of something currently unknown. No, we cannot rule out that AI completely changes the competence profile needed by many university graduates to flourish and contribute to society. But we also do not have a positive indication that this might be the case. So, for the time being, I suggest we can safely ignore uncertainty two. It is not of the kind that should worry us as educators, until such time as we have a positive reason to do so – as well as at least a vague idea where the journey could go. The first uncertainty, hower, is real. Should we encourage students to integrate AI into their study work, or should we discourage it?

Addressing the Proposal by Pascal's Wager

I suggest that the situation and problem just described can be addressed using a method of thinking invented in the seventeenth century by a Philosopher called Blaise Pascal.

Before taking a couple of pages out of this fascinating minds' book, let me clarify what kind of AI use by students I think can be considered from a Pascal's Wager perspective. It is the use of AI to solve tasks, not the use of AI to create tasks (such as mock exam questions).

Pascal's wager:

Either god exists and there is a final judgment or there is none.
We cannot know, but we must choose.
If we choose to live as if there was a god and there turns out to be none, then we lose nothing. We live a virtuous life and are good there.

If we choose to live as if there was no god, however, then we will be very, very miserable if we spend eternity in hell.

              God exists (p)   God does not exist (1-p)   ------------------- ---------------- --------------------------   Wager for God       infinity         f1   Wager against God   f2               f3

The core idea of the set-up is to see that one choice – wager for God – strictly dominates the other: in the scenario where God does not exist but we still wagered for God, nothing bad happens (Pascal suggests that being a virtuous person and a good friend is not the most horrible life imaginable). However, if we wager for God and it turns out that God exists, even if the likelihood of that scenario is very small, it gets outweighed by the fact that the utility of that scenario is infinite. So, if we are rational, we should wager for God, even if the likelihood of God's existence is very small. This option strictly dominates.

Now how can this apply to AI in Education?

Either we still teach the hard, slow way to become good at coding or first-order logic or text analysis, or we think that this is not necessary anymore, let students use AI for that and also grade them accordingly. See the following table

                                                         Students do need the hard way (0.5)   Students don\'t need the hard way (0.5)   ---------------------------------------------------------- ------------------------------------- -----------------------------------------   Wager for teaching students the hard way                   10                                    -10   Wager for teaching students the \"easy\" AI enhanced way   -100                                  50

Now, in analogy to Pascal's wager, there seems to be one scenario that dominates: The scenario where we stop teaching the hard way, but it turns out that students still need it. I think this is the only scenario that is clearly unacceptable, and that all teachers are required to work towards avoiding. The other scenario where our wager goes wrong, where we teach students the hard way but it turns out that they could have acquired the necessary skills and virtues also with AI is clearly less bad. They still end up as able, qualified graduates, ready to make a mark, but it turns out that there would have been an easier way.

So, in sum, it seems that we can confidently opt for continuing to teach the students the hard way, until such time as there is really very, very solid evidence that they do not actually need this. It worked for pascal, it works for us.