In January of 2022, I described my first attempt at ungrading as an amuse-bouche , a palate cleanser. I needed to strip everything away. I needed distance from points, averages, and the familiar machinery of grading so I could think clearly again about learning, power, and my role in the classroom.
At the time, that framing felt right. I needed to disrupt myself.
Now, looking back at my earlier posts, I realize something important. I cleared the palate, but I never fully designed the meal.
This post is not a rejection of ungrading. It’s an attempt to name what that first experiment revealed, what it failed to resolve, and why I no longer believe that simply “removing grades” is enough.
Ungrading Helped Me Decenter Myself, But That Was Only Step One
My early writing about ungrading was animated by one central goal: decentering the teacher.
I wanted to interrupt the power dynamics of the classroom. I wanted students, particularly future teachers, to question what counts, who decides, and why assessment so often feels punitive rather than formative. I wanted them to see grades not as neutral measures, but as cultural and political tools.
In many ways, ungrading worked.
Students asked better questions. They reflected more openly on learning. They problematized schooling itself.
However, something else also happened. Something I named at the time but didn’t yet fully comprehend.
Students were anxious. They kept asking for permission. They apologized constantly. They didn’t trust the system, even when I told them it was theirs.
I remember writing:
“For some reason, I had a larger than normal number of students coming to me to provide excuses… This message didn’t get through for most students.”
At the time, I interpreted this as conditioning. Years of schooling had trained students to perform for grades, and it would take time to undo that.
That explanation wasn’t wrong, but it was incomplete.
The Hard Truth: Ungrading 1.0 Was Still a Game
In 2022, I wrote:
“Assessment is ultimately a game… By deciding you want an ungrading or alternative assessment system in your classroom, you’re deciding not to play the game.”
What I didn’t fully see then was that I hadn’t stopped the game. I’d just changed the rules.
Declaration quizzes, self-assessment checklists, flexible deadlines. These removed numbers , but they didn’t remove performance. Students were still trying to figure out what I wanted. They were still managing impressions. They were still asking, implicitly:
“Am I doing this right?”
For some students, that ambiguity felt freeing. For others, it was destabilizing.
Freedom without structure doesn’t feel like freedom to everyone. For many students, especially those who have learned to survive school by decoding expectations, it feels like risk without a safety net.
In trying to decenter myself, I had also removed something else. Infrastructure.
The Anxiety of Freedom Is Real
One of the most consistent pieces of feedback I received, both from students and colleagues, was about anxiety.
Some students didn’t want to judge their own work. Some didn’t trust themselves to do it “right.” Some felt abandoned rather than empowered.
I used to think this was a form of resistance to change. Now I see it more clearly as a rational response to uncertainty.
Ungrading asked students to take responsibility for their learning, but it didn’t always provide them with shared, visible structures for how that responsibility could be enacted. I removed judgment, but I didn’t always replace it with something sturdy enough to hold people up.
That’s on me.
Then Came AI, and It Made the Cracks Impossible to Ignore
To be clear: generative AI did not create these problems.
I was already writing about equity, bias, and unidimensional assessment years before ChatGPT entered the classroom. I was already uneasy with compliance-based schooling. I was already questioning what we were really measuring.
But AI acted as an accelerant.
When a machine can summarize readings, generate discussion posts, and answer quizzes more quickly and accurately than most students, the compliance game is compromised. The question is no longer “How do I stop cheating?” but:
“Why am I asking students to perform tasks that are so easily automated?”
AI didn’t break assessment and evaluation. It exposed how fragile it already was.
And it forced a realization I had been circling for years:
We can’t just remove grades. We have to redesign the architecture of learning.
From Assessment as Judgment to Assessment as Infrastructure
This is the shift I’m trying to name now.
Ungrading helped me stop judging. However, students still needed something to orient themselves to see progress, understand expectations, and feel grounded.
What I’m working toward now is not assessment as evaluation, but assessment as infrastructure :
- Structures that make learning visible
- Systems that support agency without ambiguity
- Shared labor that replaces individual performance
That realization has led me to my current approach, which I’ll describe in the next post. Not as a solution, but as the next iteration.
Ungrading wasn’t the destination. It was the clearing. The work now is building what comes next.