Last week, I participated in a panel on Education and AI at the Faculty of Arts and Social Sciences Teaching Day. The panel was led by Danny Liu from the Education Innovation Team. The other presenter was Susan Thomas from the Discipline of English and Writing. Susan presented on “AI and the New Colonial World Order”. I presented “Some Reflections on Generative AI and the Value and Nature of Educational Goods”. Here’s a transcript of my presentation:
We sometimes think that part of the point of teaching and education is that our students might acquire certain kinds of knowledge, skills, and dispositions, that is, that they might acquire certain educational goods.
We sometimes even try to design assessments to determine whether students have acquired the relevant kinds of knowledge, skills, and dispositions, we think they should acquire.
Do recent developments in generative AI make all of this a little pointless? (Or might not too distant developments do so?)
I don’t mean that recent developments in generative AI make traditional kinds of knowledge, skills, and dispositions irrelevant. They will make some kinds of knowledge, skills, and dispositions irrelevant. But many traditional kinds of knowledge, skills, and dispositions will remain relevant.
What I mean is: do recent developments in generative AI make the acquisition of traditional forms of knowledge, skills, and dispositions, irrelevant? If so, they’d make assessing to determine whether students have acquired the relevant knowledge, skills, and dispositions, pointless.
My view is that, contrary to what we as educators would like to believe, recent and future developments in generative AI will make the acquisition of some traditional forms of knowledge, skills, and dispositions irrelevant, and not just because it will make some forms of knowledge, skills, and dispositions themselves irrelevant.
But I think that it won’t make the acquisition of all traditional forms of knowledge, skills, and dispositions irrelevant.
There are two ways that recent developments can make the acquisition of certain kinds of traditional educational goods irrelevant.
The first way is obvious, and it relates to the fact that the value of an educational good can depend on its setting.
Consider the skill of touch typing. There was a time when this was a valuable educational good. It was valuable as a means to finding employment. It is no longer so valuable. Recent developments in generative AI are clearly going to have an impact on the value of some traditional educational goods of this kind.
The second way that recent developments can make acquisition of certain kinds of traditional educational goods irrelevant relates to the fact that the possession of an educational good can constitutively depend on its setting.
This fact isn’t as obvious as the previous one. This is not so obvious, because we tend to think of knowledge, skills, and dispositions, as “intrinsic” things, things inside our heads that we carry around wherever we go. But this isn’t quite right. What skills and dispositions we have, and even what knowledge we have, may depend on our setting. My knowledge of other people’s phone numbers is much better when I have my phone on and its battery is charged. My ability to spell is much better when I remember to turn on the spell checker.
We are used to thinking that physical abilities can constitutively depend on external things. But the same is true for cognitive skills and dispositions. What knowledge, skills, and dispositions we have may depend on our setting.
To see why this matters, return to my opening point about how we think that part of the point of teaching and education that our students might acquire certain kinds of knowledge, skills, and dispositions. We tend to think of this in terms of getting things into the heads of students. But as their circumstances change, they may simply come to acquire “knowledge”, “skills”, and “dispositions” by other means than getting them into their heads. This is clearly another way that recent advances in generative AI can make this model look irrelevant.
To the extent that anyone needs specialised knowledge, skills, and dispositions, it may simply be those relevant to interacting with generative AI in the relevant ways to count as having the knowledge skills and dispositions we used to think were important.
There’s a “brave new world” interpretation of the “two-lane” approach that has been proposed as a response to recent advances in generative AI at the university. That interpretation would be based largely on a response to these considerations that concedes the pointlessness of acquiring traditional kinds of knowledge, skills, and dispositions. The specialised knowledge, skills, and dispositions that need to be “acquired” and assessed for under “assured assessment of learning” would just be that necessary to count as having the knowledge skills and dispositions we used to think were important in the context of “human-AI interaction”.
Of course, many of us would think that this is too concessive. And I am not suggesting that the “two-lane” approach is in any way intrinsically tied to this interpretation. But this interpretation certainly focuses the question: what would be lost if we adopted this interpretation? Why might the acquisition of some traditional forms of knowledge, skills, and dispositions, matter? Why might it matter that some forms of knowledge, some skills, and some dispositions, not be “off-loaded” in the way suggested by this interpretation?
Put like this, the question brings some traditional liberal concerns about freedom, autonomy, not being dominated by others, not being vulnerable to others, not being subordinated to others—values traditionally associated with “liberal education”—to the fore. (Side note: these are the values that the BA was built around, but it seems that we’ve lost sight of this).
This way of putting the question highlights the importance of the skills and dispositions of a critical and active citizenry to certain conceptions of human freedom and autonomy in ways that our usual facile appeals to “critical thinking” do not.
Some of the things that those of us in the arts, humanities and social sciences, think it is important that students acquire as part of their education, some skills and dispositions, are such that they cannot be “offloaded” without undermining their value as educational goods. Or, so I claim without much argument here.
If we want to be able to defend a non-deflationary answer to the question “what’s the point?” when it comes to what we do, we had better find ways of articulating and defending an answer along these lines.