My AI Teacher Bestie Needed a Devil’s Advocate

I spent about 36 hours going back and forth over half a class period.

My medical imaging lesson was running long. We were exploring the question, “How can medical imaging help us understand injuries and healing?” I still had work to finish, a mini lesson on structures, systems, and subsystems that I wanted to teach, and a possible activity where students would critique arguments about medical imaging. I also had a calendar, another lesson waiting, and one flex day that I could use now or save for later.

And because apparently that wasn’t enough to think about, I was also worried about the notebook page. We already had materials together on one page. Should I squeeze the remaining work into that space? Start a new page? Give the second day a new question even though we were still working on the same one?

This is the kind of decision that can take up a surprising amount of space in my teacher brain. I could make a reasonable case for slowing down. It’s early in the year. Students are learning routines. The connections matter. I could also make a reasonable case for picking up the pace. Every extension uses time, and there are other concepts that deserve their turn.

Eventually, I brought the whole dilemma to AI and asked for some food for thought.

The first response was supportive. Start a new page. Give the learning some room. If the systems mini lesson and argument critique have a clear purpose, a second day could be worthwhile. Use the flexibility available to you.

It was a thoughtful response, and I could feel how easy it would be to accept it. So I asked for something else.

“You sound like a supportive teaching bestie. Can you give me the administrator telling me to push pace and content response?”

That changed the conversation. The next response asked me to identify the essential learning, finish within the first half of the period, and move forward. It questioned whether the additional activities needed to happen immediately. It also reminded me that small extensions can add up across a unit.

The point that stuck with me was whether the learning actually required another full day, or whether I could build a worthwhile full day around what remained. I certainly had enough ideas to fill the time. That didn’t automatically settle whether I should use it.

Then I asked for one more perspective: the friend outside education.

That response helped separate the decisions I had bundled together. The notebook could get a new page without the lesson getting another full period. I could plan to finish in half a period and still change course if students showed me they needed more time. I didn’t have to decide every possible outcome before class started.

By then, I had heard three different ways of looking at the same situation. Each brought something useful into focus: making room for learning, protecting time for upcoming content, and keeping a manageable decision from becoming an all-or-nothing choice.

I had asked for those competing perspectives on purpose. I wanted to hear the strongest case for different choices and see which parts held up against what I knew about my students and my classroom.

That also made something clear about this use of AI. It could make slowing down sound convincing, and it could make moving ahead sound convincing. A confident response alone couldn’t settle the question. Asking it to play an administrator gave me an argument to consider; it didn’t give it firsthand knowledge of my classroom. I still had to examine the advice and decide what applied.

There was another piece I needed to give myself permission to consider: I teach everything from resource science to honors science. I had been trying to settle on a plan for “the lesson,” but this could also be an opportunity to differentiate. I told myself that not all of my classes need to be evaluating arguments yet. Some are ready for that. Some could work on making a claim and selecting evidence to support it. Others are still working on reading fluency and identifying the main idea in the text.

The medical imaging question can stay the same while the work changes. One class might examine whether an argument is supported by the images. Another might select an observation that supports a claim. Another might spend more time reading and making sense of the information before using it. That gave me another way to think about pacing: what is the most useful next step for the students in front of me? I can build toward evaluating arguments while giving students practice with the pieces they need along the way.

My decision was to try to push the pace, have the next lesson ready, and keep a fallback plan available. I would give myself a target instead of automatically slowing down, then adjust the task and the time based on what students in each class showed me. At the time of writing, that is the plan. I still have to teach the classes and find out how it goes.

I think this is a useful way to approach some of the everyday decisions that follow us home. We can ask AI to help us examine an idea from several angles, especially when we notice that the first response feels a little too comfortable. A supportive colleague, a skeptical administrator, a friend outside the profession, or a thoughtful devil’s advocate can each be a useful perspective to request.

For someone wanting to try it, I would start with a prompt like this: “Here is the decision I’m wrestling with, what I’m leaning toward, and the constraints I’m working under. Give me the strongest reasonable case for my plan, then the strongest reasonable case against it. What information would help me decide?”

For anyone keeping score, yes, AI helped me think this through and write about it. I even asked it to disagree with me—something the internet usually provides for free. I’m still the one teaching the classes and finding out whether the plan works.

I could have kept turning this over for another evening. Instead, asking for a few different perspectives helped me choose a next step. Tomorrow, I’ll walk into class with a plan, a stopping point, and some room to adjust. That feels like a useful outcome for a conversation that started with whether I should turn the notebook page.