Grumpy Rumblings (of the formerly untenured)

CG asks:

I’m teaching a new undergrad seminar this fall that meets in one 2.5 hour block per week. I need to figure out how to A. keep them from using AI for everything and B. keep them engaged for that long chunk of time. I might do some in-class writing prompts. I’m also going to ask them each week to look up some stuff relevant to class, post two sentences about it on Canvas, and bring it in to share. This is a subject that the people who take it tend to be very personally interested in so I think that second tactic will be effective.

Any clever ideas from fellow academics to help me as I gear up for this course would be greatly appreciated!

AI

I think the first thing regarding AI is to think about what the purpose of the class is.  If the purpose is to learn material, then AI doesn’t have as much of a place because it will substitute for learning (though see comparative assignments below).  If the purpose is to create a product applying material already learned, then AI might have a place, but you will need to define it and they will need to demonstrate how they use it, and importantly, how they check it for accuracy.

A lot of my colleagues have been incorporating AI in their upper-level classes by having assignments in which they request that the assignment be done with AI and then they request it be done traditional ways and either the student or the class as a whole has to compare the two processes and outputs.  AI is good for some things and isn’t good for other things.  Some assignments have different groups doing the same thing in separate teams, each with a different AI or no AI at all, then they come together to compare outputs (right now it seems like the big learning thing is that AI is faster but still needs a lot of oversight, which is an important lesson).  Some assignments are individual where they do both the AI version and the real version.  (But be careful with these– they must be novel assignments– if it’s from a textbook then AI will get things more correct than if it’s something new.  Still, students tell me that they’re shocked by how much AI still just makes things up.)

Some of my colleagues have gone back to flipped classrooms where the lecture happens outside of class and all of the assignments happen within class.  That way they can actually watch people not using AI.

Keeping students engaged

This one I have a lot more success with because most of my classes, *including* math classes, meet for 2.5 hours once a week, so I have had a lot of practice.  (Not this semester, but most semesters that’s how it has worked out.)

The big thing is that you have to mix things up.  You cannot straight lecture.  You can do interactive lecture where you ask a lot of questions to check for understanding or to involve the class in lecture creation.  Basically if it’s a lecture class you lecture, check for understanding, have them do a problem, and repeat.  Sometimes there will be class-relevant games or Stata exercises to break things up.

In seminars I like to provide discussion questions with the reading, which I have them turn in as a completion grade.  Then, importantly, we go through the discussion questions — sometimes everyone will answer the same question and we will compare and contrast answers.  Sometimes I have each person answer a different question.

I guess generally for discussion classes for me, most of the prep-work comes up front– I figure out what all the assignments are going to be and then I use their answers to those assignments to actually make class happens.  Students say they love the way that they never feel like they’re just doing busy work– everything they do ends up being used in class.  We actually visit the readings and go through the questions they were assigned.  They actually post the information they found to the class.

One thing I really like to do is give them assignments where each of them has to bring in information about a different group, for example, voting rules for different states.  Then we combine all of that information on the board and see what big picture things we can take away from comparing and contrasting.  One really illuminating class we looked at privatization of I can’t remember if it was prisons or child welfare or both, but what came out of a dozen different articles that students found (this is NOT my area of expertise and the syllabus needed to be upgraded, so I told them to each find a paper and post it as soon as they found it, no two people could have the same paper) was that privatization can work and make things more efficient and not more dangerous if there’s good governmental oversight, but if there isn’t good oversight it’s dangerous.  (Don’t quote me.)  Those points where we learn things are new to even me seem really special.

An assignment I stole from my friend is a bluesky assignment where people have to follow some number of (in our case) economists, and then they post the tweet they found most interesting for the week, and then read everyone else’s tweets and we would go around the room and each person would say which tweet of those we found the most interesting and we’d discuss them and how they related to the class material.  (Usually multiple people found the same one the most interesting.)  There are early deadlines for each step of the process as well so they can’t just do the classwork the night before it is due and then fall asleep in class.

In class writing is good.  In class problem solving is good.  Conversations are good.  Guest lectures are good– it is amazing how you can get local government people to zoom or even come in to give a talk about what their program does.  You can also get policy-makers and researchers to zoom in and give a talk.

Grumpy Nation Teachers and (former) students:  What advice do you have for CG?