In my previous blog post on the Microsoft Research study about GenAI and expertise I ended with a troubling realization that GenAI may not be the best options for learners. As I wrote “This analysis raises particularly thorny issues about AI use in education. If expertise is a key prerequisite for effective AI use, then learners – who by definition are not experts – are in a particularly vulnerable position.”

To give some context, I introduced, in that post, a framework with four scenarios for AI use based on the interaction between domain expertise and AI literacy, as captured in the image below.

If you look at the 4 quadrants it becomes clear that the most vulnerable position is at the bottom-left space (which I called the “Novice’s Dilemma”). These are people who neither have domain expertise nor knowledge of AI. In an educational context, it means that there is a population (learners) who are doubly disadvantaged—unable to evaluate AI’s outputs for accuracy and unaware of when and how AI might lead them astray. This makes them particularly susceptible to accepting AI-generated content uncritically.

This, of course, creates a problem for education and educators. If expertise is necessary for effective AI use, then students—who by definition are not experts—are in a precarious position. The standard solutions (exhorting users to “think more critically” or providing verification checklists) seemed inadequate, especially for those who need them most.

I left that post with an unresolved question: How do we help learners move out of this vulnerable position?

One solution that has been offered is for us to see GenAI as being a “teacher’s tool” rather than one for learners. That seems somewhat inadequate to me, for a range of reasons (which may be a future blog post).

A different answer came to me while recording the latest episode of AIR|GPT, a monthly podcast where a bunch of us “ed tech types” get together (supposedly in an airport lounge) to chat about whatever is on our minds at that time. No surprise, the focus of our latest conversation was on the Microsoft study. While I was listening to my co-hosts comment on the double jeopardy that learners face, that the answer hit me—an answer so simple, so fundamental that I felt almost foolish for not seeing it immediately:

The key element that moves learners out of the novice quadrant is CARING about what they’re learning.

This is quite definitely NOT a new insight. Educators have spoken about it forever. My friend and colleague Yong Zhao, with whom I co-host the Silver Lining for Learning podcast, has been making the same point for years in his critique of standardized education and advocacy for student-driven learning.

In rediscovering this fundamental truth about learning, I’m reminded of T.S. Eliot’s observation:

We shall not cease from exploration, and the end of all our exploring will be to arrive where we started and know the place for the first time.

When students genuinely care about a subject, the motivation to critically evaluate information arises naturally—not as an abstract academic exercise done to please teachers, but because the subject matters to them personally. This act of caring, transforms the cognitive effort of verification from a burden into meaningful engagement.

This reminded me of what Dewey called the four “primary impulses of learning:” Inquiry, Construction, Communication and Expression. He characterized these impulses as “the natural resources, the un-invested capital, upon the exercise of which depends the active growth of the [learner].” He posed a critical question for educators: “What are we to do with this interest—are we to ignore it, or just excite and draw it out? Or shall we get hold of it and direct it to something ahead, something better?”

Incidentally, these four impulses were foundational to how we conceptualized our EDT180 Technology Literacy Course at the College.  Rather than organizing content around specific technologies (which quickly become obsolete) or abstract “digital literacy” skills, we structured learning experiences around these four natural impulses. I have written about this experience elsewhere: Embracing failure in a first-year technology course.

In this particular context, (of the use of GenAI in education), I would argue that the ideas play out in the following manner:

Inquiry: When students pursue questions they genuinely care about, they naturally become more discerning about sources and the act of making meaning. Key of course is a focus on “discovery” curiosity (as opposed to “deprivation” curiosity)—something I have written about in a post titled: The darker side of curiosity.

Communication: Preparing to share ideas with real audiences motivates students to scrutinize content carefully and the social dimension creates natural accountability that drives critical engagement with AI outputs.

Construction: Building something meaningful creates ownership. Students invested in their creations validate AI contributions more carefully because the quality of what they’re making actually matters to them.

Expression: When students express themselves through artistic and creative media, they’re bringing their unique aesthetic sensibilities and personal identity to their work.  This naturally motivates them to ensure any AI assistance enhances rather than diminishes the authentic expression of who they are and what they value.

This perspective of caring for what we learn, driven by these primary impulses, helps us reframe how we think about the expertise paradox of GenAI. Rather than seeing the expertise requirement as an insurmountable barrier for novices, we can view caring as the bridge that helps them develop both domain knowledge and AI literacy simultaneously.

And this actually helps explain the somewhat paradoxical result in the Microsoft study: where experts working with AI experienced increased cognitive load rather than decreased effort. I would argue that this was precisely because the experts cared enough about quality outcomes to do the extra work of verification and curation. When students have real stake in learning something they willingly take on this cognitive effort not as a burden but as a natural part of pursuing something meaningful.

The rush to integrate AI into education has often focused on technical skills, verification strategies, and abstract critical thinking. These aren’t wrong, but they miss the fundamental motivational foundation that makes such skills meaningful to learners.

We’ve been looking for complex solutions when the answer is both simpler and more profound. As Dewey understood a century ago, learning is driven by these natural impulses that make knowledge meaningful and worth pursuing. The question has always been:

Not “What?” but “WHY?”

The implications for education in the age of GenAI are actually quite straightforward. Rather than starting with AI verification strategies, we should first create conditions where students develop genuine interest in what they’re learning. That often means starting from what they care about. It also means designing experiences around Dewey’s natural impulses—authentic inquiry, meaningful communication, creative construction, and personal expression. In this context GenAI is NOT a shortcut for completing relatively meaningless assignments with minimal effort. Perhaps most importantly, we must recognize that critical evaluation naturally flows from caring; when students are invested in the quality of their work, they develop the discerning stance needed for effective AI use without requiring elaborate verification frameworks. Plagiarism in this context is not a relevant concern.

I would be the last to suggest this shift would be easy. The grammar of schooling has become deeply ingrained in our educational systems, our students’ expectations, and our own practices as educators. I’m experiencing this challenge firsthand in the creativity course I’m teaching. When I told students, “If you’re doing something just for me, let me know and I’ll cancel that assignment,” they initially couldn’t believe it—so discordant was this approach with standard academic expectations. Yet despite the difficulties, fostering genuine caring remains the most powerful solution to the novice’s dilemma.