Originally Published May 15, 2024
As I was walking the halls the other day, I overheard a student interviewing another student and asking a question that broke my heart. She asked, “How do you deal with the stress of education, especially the stress of finals?” Learning should be joyous, not stressful. It’s no wonder that most people run away from school as fast as they can. We naturally flee stressful situations.
It should come as no surprise that our students are stressed. We have normalized stress as part of the educational process. Students expect to be boxed on the assembly line of their future. Assessments are there to make sure they are conforming to the process.
Education is a linear process. A myriad of steps that trace from kindergarten to a PhD measure progression. It’s an assembly line for minds. The problem emerges when the factory becomes more important than the product.
This is not how learning works and not how the real world looks. We persist in this fiction because it is easy to wrap our heads around it in the abstract. Completion and grades become the focus. In my book, Learn at Your Own Risk, I explore how the currency of grades undermines learning. Without assessments, you don’t have grades.
Generative Large Language Models inject yet another element of chaos into this system. Last year, students seized these tools as a way of attacking the arbitrary linearity of their academic confinement. Some call this cheating, but what are they cheating, their learning or the factory line?
The primary purpose of assessments these days seems to be as a quality control measure for advancement up the assembly line that results in graduation. We use them as gatekeepers to the future because we view graduation and the legitimacy that it confers as the stepping stone to employment.
This is a perversion of their intent. It obscures the different purposes which assessments can serve. In a system where we measure forward motion with a linear advancement tool, assessments provide an extrinsic motivator for students to focus on whatever material we expect them to remember and then immediately forget half of it or more.
Often these are completely arbitrary in the eyes of the students, and therefore an unfair judgment of their abilities. If this is how students perceive assessments, it’s no wonder “cheating” is endemic.
Most assessments emphasize conformity over creativity. They try to push everyone to a normative ideal. Life is full of arbitrary cruelty, but in few places have we institutionalized this more so than in the world of academic assessment.
We can discard summative assessment in favor of formative assessment. However, it’s much harder to compare students against an arbitrary pecking order if you do. If you are interested in figuring out what students can do, judging them relative to other students is probably the worst way to do that.
If we based our method of assessment on giving the student insight into what they can and cannot do, it completely changes the dynamic of the learning process. Pride of accomplishment replaces fear of failure.
Learning is impossible without failure and yet linear summative assessment routines teach students to fear failure. In this kind of world, students have grasped generative AI to cheat failure in a system that doesn’t really benefit them anymore.
It’s a natural reaction. Integrity and faith in the system only matters if the system is worth defending.
Education has to mean something if it hopes to overcome the technological debasement of its arbitrary processes. The only way to overcome this is if we give the students a mechanism for creating meaning instead of jumping through hoops.
AI is an augmented creativity tool. We need to use it as such to create and execute meaningful assessments.
Traditionally, education is at its best when it gives people meaning and context to explain and organize the randomness of the world. We’ve lost that when our processes are as arbitrary as the world we are trying to explain.
It is hard for anyone to overcome this. The entire system works against learning and whatever assessment practices don’t conform to the norm. Sometimes this is explicit in institutional assessment mandates. Sometimes it’s very passive because students rebel against assessment techniques that turn their lives into jumping through hoops that they find hard to understand.
As I have written about before, I practice unorthodox assessment mechanisms in my class. I spend most of the class teaching the students how to learn. I do this by gradually asking them to do more and more complex tasks. Students are assessed through completing the assignments.
The process of students struggling with the tasks is where the learning takes place. Those who struggle and overcome get the better grades. Those who give up or don’t engage in the struggle fail.
None of these tasks operate in isolation, and like in a real-world project, they build upon each other. The grading system is cumulative, not deficit-based. I use points instead of averages to calculate how my students are progressing through the class.
This semester, for the first time, I’ve integrated generative AI into the process. Students start out by using a chatbot to understand the problems they are investigating in the class. We then workshop that exercise as part of the process.
The one thing I’m not able to avoid, however, is the overriding concern the students have about failure. I don’t have any high-stakes assessments in my class. As a matter of fact, there are no tests at all. Failure on any assignment doesn’t doom them to failure at the end of the class. Unfortunately, the system has trained them to view assessments in this manner.
And this is where the system comes back in. I am lucky in that I do not have arbitrary assessment requirements from my institution. I have taught elsewhere where random multiple-choice tests are required to assess the students’ mastery of content. This would break the spirit of my class entirely.
But there is a larger constraint that I cannot avoid, and that is the insularity of the common academic schedule. During an academic term, you were taking government or English or history or math and that’s what we’re going to talk about, which ends at a fixed time.
In order to fix assessments, we’re going to have to do a much better job with interdisciplinarity and to view the overall learning process much more fluidly. Knowledge is not confined to one class. One conceit that AI exposes is the cross-disciplinary reality of knowledge.
As long as we separate college (and secondary school) into little boxes that bound learning, we will have small demeaning assessments because those boxes’ primary connection with one another lies in the arbitrariness of grades.
Grades matter to the system. They only matter to the students because they matter to the system. Technology breaks that.
College used to produce well-rounded individuals who could handle a variety of nonlinear challenges. Our current systems of assessment work against that outcome.
Assessment should encourage exploration, not constrain it. We can, and must, develop tools and systems to fix this if we want to graduate students capable of handling the wicked problems of today and tomorrow.