What If AI Literacy Is Really Learning Literacy?

We talk a lot about what students need to know about artificial intelligence.

They need to understand what these systems can and cannot do. They need to recognize bias and misinformation, think about privacy and ethics, and learn how to prompt effectively. Schools are also beginning to draw boundaries around appropriate use: when AI is allowed, when it isn't, and which parts of an assignment still need to belong to the student.

All of that matters. But lately I've been wondering whether we're spending so much time defining AI literacy that we haven't spent enough time asking where that literacy is ultimately supposed to take the learner.

What are we trying to help students become?

Recently, I spoke with students who had participated in a national student senate focused on developing recommendations for AI in schools. What struck me wasn't simply what they thought the rules should be. It was how naturally some of them talked about their own learning.

One student described a distinction in math. After completing a study guide, she might ask AI to generate additional problems with different numbers so she could practice more. But asking it to complete a take-home test or corrections was different. In one case, AI was creating another opportunity for her to do the learning. In the other, it could remove the very work through which the learning was supposed to happen.

Their test for that distinction could be surprisingly simple: Did I actually learn anything?

Not long afterward, I had a very different conversation with learning researcher John Hattie. We talked about students becoming question askers rather than simply question answerers, recognizing when a learning strategy isn't working, critiquing what AI produces rather than simply accepting it, and ultimately knowing what to do when they don't know what to do.

The connection between those two conversations has stayed with me. Both point toward a conception of AI literacy that has less to do with mastering a technology and more to do with understanding your own learning.

Using AI and leveraging AI aren't the same thing

I've gradually learned to use AI to help organize my thinking. I tend to generate ideas, make connections, and see possibilities fairly easily. Organizing all of those thoughts into a coherent structure can be harder for me.

AI can help me see that I've actually got three different ideas tangled together. It can challenge a connection I'm making, organize a messy collection of thoughts, or help me see a structure I haven't yet articulated. But I don't necessarily want it to generate the idea I'm trying to develop. Sometimes I want to wrestle with that myself.

What's interesting to me isn't that an AI system can organize information. It's that, through using it, I've gotten better at recognizing where the technology is useful to me and where I don't want it to intervene.

That boundary isn't fixed. I've learned to recognize it through use.

A novice user can learn how to write a prompt. A more sophisticated user begins to understand what to ask AI to do. But perhaps an even more sophisticated learner understands why they are asking it to do that particular thing at that particular moment.

That requires knowing something about the technology. It also requires knowing something about yourself.

The learner may eventually need to make the decision

Right now, much of the conversation in schools understandably focuses on adults determining the boundaries. Teachers decide whether AI is appropriate for an assignment. Districts develop policies. There are good reasons for that, particularly while students are still developing foundational capabilities. But that can't be the final destination.

Eventually, students leave the classroom. There won't always be a teacher standing beside them saying: use AI for this part, but not that one. Ask it for feedback, but don't ask it for the answer. You know enough now to use the calculator. You don't know enough yet to outsource this step. The learner will have to make those decisions.

Doing that well requires much more than knowing how to operate AI. A learner has to be able to ask: What am I actually trying to learn? What do I understand already? Where am I stuck? Is this something I need to work through myself?

And perhaps most importantly: Am I understanding more, or am I simply producing something better? Those two things can look remarkably similar from the outside.

A student may submit a better essay without becoming a better writer. Another student might use AI during the writing process and encounter arguments she hadn't considered, critique them, reorganize her own thinking, and ultimately understand the subject more deeply. The presence of AI doesn't tell us which occurred. Neither does the quality of the final product. Increasingly, the learner needs to know the difference.

Maybe AI literacy eventually becomes learning literacy

There will always be knowledge students need about the technology itself. Schools will continue to need boundaries, particularly as these tools evolve. But perhaps the deeper capability we're trying to develop isn't simply proficiency with AI. It's agency over one's own learning.

That means developing a student who understands what learning feels like; who recognizes when understanding has deepened and when it hasn't; who can change strategies when something isn't working; who knows when to ask for help, what kind of help to ask for, and when assistance is beginning to replace something they need to develop themselves. That student can use AI differently.

They aren't simply a consumer of whatever the machine produces. They can question it, critique it, redirect it, ask it to expose weaknesses in their thinking, use it to create opportunities to practice, and hand over some work while deliberately holding onto other work. In other words, the destination isn't a student who has memorized the rules for responsible AI use. It's a student increasingly capable of making those judgments for themselves.

Which brings me back to the students I spoke with. They were already asking a remarkably sophisticated question: Did I actually learn anything?

Maybe that's not merely a good rule for responsible AI use. Maybe it's pointing toward what AI literacy is ultimately for.

The destination of AI literacy may not be a student who knows how to use AI. It may be a student who understands learning, and themselves, well enough to know how AI should participate in it.

Continue the conversation

If this article sparked a new way of thinking, you'll find more original perspectives, conversations, and insights at InflectionED.com.

Subscribe and listen to my full conversation with Dr. John Hattie on Education Talk Radio, releasing September 25, 2026.