Stop Assessing Students Like AI Doesn’t Exist

The problem isn’t AI. It’s what we’ve been measuring all along.

by Jim Bowler, EngAIged

For years, my classroom rhythm was familiar: teach the lesson, assign the homework, give the test, score the answers, move on.

Imagine I’m back in my classroom, teaching photosynthesis. I ask my students to read about it at home, complete an assignment, and come to class. A few days later, they take a test. If they score 9 out of 10, I assume they learned it.

But what exactly did I learn from that score?

• Did they understand photosynthesis?
• Did they really understand the relationship between sunlight, water, carbon dioxide, and plant growth?
• Could they apply that knowledge to a new situation?
• Could they explain why one plant grows while another struggles?

Or did they memorize enough to answer my questions?

This issue has always mattered. AI just makes it much harder to ignore.

AI Changed the Evidence

My students can now use AI to generate an essay, answer questions, solve a problem, or create a polished response in seconds. Most students with a basic grasp of AI could earn an A on my take-home assignments.

The finished work my students hand in is no longer a reliable proxy for learning.

AI didn’t create this problem. In my years in the classroom, I watched students memorize without understanding, copy from classmates, get help from parents, and turn in work that looked better than what they actually knew. AI simply made the gap between the product and the learning much larger.

Of course I can ban the use of AI in my class. I can run AI-detection and plagiarism software. But that’s a losing battle, and it doesn’t prepare my students for a world where AI is an everyday tool. (See AI Bans Don’t Work)

That doesn’t mean students no longer need to know facts. It doesn’t mean quizzes are useless. And it certainly doesn’t mean assessment should disappear.

The Quiz Still Matters

I’m not throwing away the short quiz. In fact, short assessments may become more useful.

Go back to the photosynthesis lesson. If I flip the classroom, my students learn the basics at home. They may read, watch a video, use an AI tutor, review class materials, or use a combination of resources.

The next day, I start class with five quick questions. That quiz has a purpose. It tells me whether my students came prepared with the foundational knowledge they need. AI can draft that quiz and score it instantly, so I know which idea the class missed before we move on. But it shouldn’t be confused with deep understanding. Here’s what I tell my students:

The quiz tells me whether you know the basics.

What you do next tells me whether you understand them.

The Real Assessment Happens Next

Now I ask my students to do something with what they know. Maybe they compare two plants growing under different conditions and predict what will happen. Maybe they design a simple experiment and explain which variable they would change. Maybe one group challenges another group’s explanation.

Now I can see much more.

• Who can explain cause and effect?
• Who can apply the concept to a new situation?
• Who changes their thinking when presented with better evidence?
• Who has memorized the vocabulary but still doesn’t understand the science?

That is where my judgment matters most.

Assessment Should Happen During Learning, Not Just After It

One of the biggest opportunities AI creates is not simply to grade faster. It is to rethink when assessment happens. Quick checks used to take enough time that I couldn’t use them as often as I wanted. AI makes them cheap enough to be routine.

Instead of relying so heavily on one finished product at the end, I can collect evidence throughout the learning process. Assessment should be embedded in it.

• Observation during an activity.
• Questions during discussion.
• A short explanation from a student.
• A revision after feedback.
• An exit ticket at the end.

The goal is not to create more testing. The goal is to make student thinking more visible.

The Obvious Objection

It’s a lot of work. I had over 120 students a day. I couldn’t watch every one of them think, and I’d have to design the activities and build the rubrics. That’s fair. But it isn’t simply more work on top of the old work. AI can score the quizzes, draft the activities and rubrics for me to edit, and help me build materials I can reuse next year.

The hours I used to spend scoring fact quizzes can become hours I spend observing student thinking.

Homework and Assessment Need to Change Together

This is also why the homework conversation and the assessment conversation are connected. If students can use home time to learn the fundamentals in ways that work for them, then classroom time can become more valuable. (See Homework Must Change)

It can become the place where my students discuss, experiment, apply, challenge, defend and demonstrate what they really understand. And that gives me better evidence.

AI doesn’t make assessment less important. It makes good assessment more important than ever. Because in a world where my students can generate answers instantly, the real question is no longer simply: Did they get the answer right?

It is: What do they understand, and what can they do with what they know?

Reader Question

“Where in your class could you watch students think, instead of only grading what they hand in?

---

How can you get started using AI to save you time and improve learning in your classrooms? Download our guide free.

---