Are we teaching AI literacy, or are we just teaching students how to outsource thinking more politely?
Too often, AI education starts with the tool. Here is how to prompt. Here is how to generate. Here is how to save time. But if students meet the AI before they meet their own thinking, what exactly are they learning? For me, the real challenge is not whether students use AI. They are and they will. The challenge is whether they can think before prompting, judge before accepting, adapt before submitting and explain before claiming ownershipAI literacy cannot be only about access, efficiency, or productivity. It has to be about agency. A student who can use AI is not automatically AI literate. A student who can question AI, reject weak outputs, improve ideas, protect their voice, and explain their decisions is much closer. So perhaps the question for AI educators is this. Are we designing AI use that strengthens human intelligence, or are we quietly designing around it?
39 replies
AI is not a cheating problem, as you have pointed out; it is an education design problem.
Prompt the human first!
I agree with the access versus agency line, because that's the real distinction. If we can't name the exact moment in the process where thinking gets protected or handed over, agency is just a value we're hoping shows up. It has to be intentional and deliberate, built into how the task unfolds, not just what the task produces.Think about it as a sequence, not a single moment. Students generate...
Agree, and the sequence point is the useful bit here, because it’s checkable. You can look at a task and see whether the order was followed or skipped, which is more than you can say for “agency” as a stated aim.The cyborg versus centaur distinction is doing real work too, but I’d probably build on on the bit at the end. Knowing which mode fits the moment isn’t a soft skill sitting alongside...
Yep, takes a critical view of assessment to move forward.
I agree with you, Michael. AI education cannot just be about teaching students how to use a tool faster. When I design learning experiences for my students, I want to make sure they still have opportunities to make decisions, explain their process, and reflect on what they accepted or rejected from AI. The same goes for assessment. If we only assess the final product, it becomes harder to see...
it’s a tightrope walk. If we’re just training pupils how to write the perfect prompt to fetch quick answers, we’re essentially teaching them how to offload their brains with better manners. True AI literacy isn't about using a chatbot to skip doing the hard graft; it's about giving them the critical distance to treat it as a flawed tool, not an all-knowing oracle. That's why I use sci-fi...
Is that AI literacy or moving into AI fluency?
Learning is dialogic, a conversation. AI flattens that dialogue.
I really like the way you framed assessment and how those different versions of checkpoints and visible learning are more important than just a final product that doesn't truly showcase their learning!