| The Educational Artifact Is No Longer Enough AI is changing education, separating what students produce from how they think.  By John Nosta |
| KEY POINTS:AI is breaking the old link between the quality of student work and the thinking behind it. Education must now consider how much students actually participate in the cognition that produces their work. As AI makes the artifact better, it’s harder to know whether the student is becoming a better thinker. |
 Source: Image by F1 Digitals from Pixabay.MIT has released an interesting new report on AI and education—what they call a “roadmap to the future.” At times, the language and perspective may feel a bit lofty and even aspirational (be humble, put humanity front and center, and teach with intentionality). Nevertheless, I think it arrives at exactly the right moment. The academic cauldron of controversy is bubbling away and the questions are familiar. What should students learn when AI can write essays and explain difficult ideas almost instantly? How should we assess learning? When should students use AI, and when should they work without it? Yes, these are practical questions, but there is something more fundamental underneath them. If education has always been partly about developing the person who thinks, what happens when some of that thinking is now being done somewhere else? The Intellectual Artifact is Changing The school has traditionally been a springboard for intellectual work. A sophisticated essay suggested that considerable thinking had occurred. A clever solution suggested understanding. A compelling argument suggested that someone had spent some time working with the ideas behind it. It was never a perfect relationship, but generally the intellectual artifact told us something about the intellectual work required to produce it. AI changes that assumption. Today, a student can produce an extraordinary intellectual artifact while contributing surprisingly little of the cognition behind it. The essay may be excellent and the argument may even contain genuinely interesting insights. Yet none of that necessarily tells us what happened inside the student. The artifact is still there and may actually be better than ever. What has changed is our ability to infer the thinking behind it. I’ve been focusing on this issue in my own writing for some time. I’ve called parts of it cognitive surrender, the curious idea of ghost thought, and more recently cognitive selfhood. They approach the problem from different directions, but they share a fundamental concern. As AI becomes increasingly capable of participating in our thinking, how do we make sure that we continue participating too? From Surrender to Participation First off, I don’t see AI as an intellectual adversary. I’ve also written about iterative intelligence, where human and machine cognition interact and the back-and-forth actually makes our thinking better. AI can challenge an assumption, expose something we’ve missed, push an idea in an unexpected direction, or make us reconsider what we thought we knew. I experience this myself, and I suspect many people who use large language models also do. But iteration can drift into surrender, sometimes without us noticing. We begin with a question and very quickly arrive at an answer, but how much of the intellectual territory between those two points did we actually travel ourselves? That seems particularly important in education because the journey has always been part of the point. Struggling with the problem wasn’t simply an inconvenient way of getting to the answer. Something was happening to us along the way. MIT recognizes part of this problem directly, particularly around assessment. The report argues, from my read, that we need to reconsider how we determine what students have actually learned in an AI-rich environment. I think that gets to something important. If AI weakens the connection between the quality of the artifact and the cognition required to produce it, then assessment has lost one of its traditional and most important signals. We may have to look beyond the work itself and ask how much the student actually participated in the thinking that produced it. In fact, they deliver an imperative.This is not an optional exercise. We must demonstrate how to integrate AI thoughtfully and deliberately into the classroom, the research enterprise, and the experience of residential education in ways that ensure learning, advance discovery, and prepare students for the future. —MIT Ad Hoc Committee on AI Use This makes cognitive participation increasingly important. We have spent considerable energy asking whether students are using AI, and understandably so. But perhaps the more interesting question is what the student is actually doing while using it. Did the student originate the idea? Challenge the response? Recognize an error? Change direction? Bring judgment to the process? Or did a prompt simply get us from the question to the solution without the friction of real learning?Those are very different cognitive experiences, even if the artifacts they produce look remarkably similar. What Happens to the Thinker? This touches something I’ve recently described as cognitive selfhood—the continued presence of the self in its own processes of thinking. Education has traditionally been one of the places where that presence develops. We struggle with a problem and sometimes discover what we think while trying to explain it. Much of that process is inefficient and occasionally frustrating. Yet I’ll suggest that some of that friction may be doing more educational work than we realized. AI will become part of education, probably much faster and more completely than we imagine. And much of that will be enormously valuable. However, there will also be moments when doing the thinking yourself has value precisely because it is slower and harder. Maybe those moments will become increasingly important because they give the student something that cannot be measured simply by looking at the quality of the finished work. And this is where I think the MIT report opens a much larger conversation. Education has traditionally been able to look at the artifact and make some reasonable assumptions about the cognition behind it. AI makes that increasingly difficult. The essay or equation can no longer serve as quite the same proxy for learning that it once did.A student may now hand us a polished artifact—a gem of intellectual brilliance. Yet we may know less than ever about what happened in the mind that produced it. So, maybe the uncomfortable (and necessary) question for education is what happens when the work keeps getting better, while our ability to know whether the student is becoming a better thinker keeps getting worse. |
| John NostaThe Digital SelfTechnology, Transformation and the Future You |
canisgallicus 