How do you know a student did the thinking, when the output looks the same either way?
Two students hand in the same task. Both pieces are clear, well structured and correct. One student worked through it. The other had a tool produce it in a minute. From the finished work alone, I cannot tell which is which.I hit this every time I plan a unit. Most of what we collect from students is a finished product, and a finished product has stopped being reliable evidence of thinking.I have a few moves I rely on, and I will share them in the replies. First I would like to hear yours.What do you build into a lesson so the thinking itself becomes visible? I mean during the work, before anything is handed in. A routine, a question you ask, a constraint on the task, something students have to show.Specific examples help most, including the ones that did not work.
14 replies
"Thinking" in the age of AI is different than what it was before AI. Now, as educators, we need to teach students to assess and improve AI output as opposed to evaluating them on the answers to questions we give them. All of my assignments now are structured in almost the same way: go ask Gemini to do this, copy the answer to the LLM: what do you think the answer is missing or how can it be...
This is the distinction between evidence of learning and evidence of performance. A polished final product might show what a student can submit, but it tells us little about how they got there, particularly when AI is involved.In my senior Physics and Chemistry classes, I have been moving towards formative checkpoints built into the learning process. Before students use AI or begin producing...
We need to find ways to Multi Factor Assess! Written assignment? No problem, but lets follow up with a podcast to demonstrate deeper understanding and critical thinking. How about students do a draft in 'exam conditions' so you can see where the gaps in knowledge are, then submit their notes, final work, version history and a reflective journal.Or allow AI to plan the assignment but the...
Mostly, I’ve found that one can’t detect thinking after the fact. One has to build it into the work, at the points where a finished product can’t reach. Here’s what I use:-Predict first. Students write what they expect and why, then check. A wrong prediction with a reason is hard to fake.-Keep the dead end. Require one approach they tried and dropped, and why. Asking for “the moment you...
Hi Damir - it's a classic problem. Homework used to be a proxy for knowing that the student had spent time on it and also that they understood it to the demonstrated degree. That obviously doesn't hold in the age of AI, where answers are cheap. In my experience, very few students turn to AI to cheat, although many of them end up that way. The four biggest reasons we see for students using AI...
Coming from an Economics teaching background, one approach I’d suggest is changing an assumption halfway through a task. A student explains why higher interest rates might reduce inflation, then you ask: “What if the inflation is mainly being driven by rising energy costs?”Before they revise their answer, ask them to explain to a partner which part of their reasoning still holds and which...
I incorporate a lot of experiential learning into my Business Studies units, from accounting simulations to managing inventory. These experiences give students something concrete to discuss (i.e. what they did, why they made certain decisions, and what they would change next time). Asking them to talk me through those choices helps me understand their reasoning beyond what I can see in the...
Yes, I co-developed a similar model as part of a global collaboration with teachers from Asia and Europe. https://humanaize-blue-classnet19.dr-harves.chatgpt.site/
Ask student what does he feel, what is his personal opinion about the topic , ask him to change somemthing and spend 3-5 minutes adjusting his assignment to it.
I like the idea of including the student's inner state and feelings in the learning process. I was just reading the study which says the Cognitive Learning Theory needs an update in order to expand and observe the non-cognitive dimensions of learning.