Here's a question that changes everything about how we think about AI in education: What if AI didn't break our assessments — it just revealed that we were never really measuring learning in the first place?

A new analysis in Inside Higher Ed makes a compelling case that for decades, we've been grading outputs — essays, problem sets, research papers — as proxies for understanding. Those proxies worked when producing them required genuine effort. But when AI makes production trivial, the proxy collapses.

Arizona State University found that 45% of their course points could be completed by AI with no evidence of actual learning required. That's not a technology problem. That's a design problem that was always there.

The educators getting this right are making one powerful shift: from asking "What did you produce?" to "Can you explain why your solution works?" That single question creates evidence AI cannot fabricate.

What's one assessment in your course that you've already redesigned — or know needs redesigning — because of AI? I'm genuinely curious what's working and what's not.

Dive deeper in today's blog post: AI Didn't Break Your Assessments — It Exposed Them → https://www.harvestkernel.com/ai-exposed-assessment-crisis/