Feedback is My Job

How far are we from exactly this?

I love LinkedIn.

Reading my connections' experiences and insights that resonate with me is priceless.

There is always at least one post or article that inspires me.

This time, it's Matthew Wemyss' latest article.

He briefly talks about how students interpret feedback based on who (or what) provides, rather than its accuracy. They have a tendency to reject feedback once they learn that it's from AI. Yet another must-read article from him.

Honestly, I still cannot believe how teachers consider using AI for assessment and feedback.

To my understanding, marking an exam paper vs having AI mark and provide feedback for the exam paper AND THEN read the feedback only are two different things.

And the latter only harms the relationship between students and their teacher.

A teacher and their students have a relationship.

Depending on that relationship, sitting an exam their teacher gave them is an opportunity for the students to show themselves; how much they know, how good they are on solving problems, etc.

And that exam paper is like a page on a diary that students and their teacher share.

When that teacher hands over the assessment and evaluation to AI, they "cheat".

Not because they don't do their job and still get that salary. No.

They cheat on their students.

They don't show the respect that exam paper deserves.

Besides, they miss a big opportunity to get to know their students' strengths and areas of improvement first hand.

No, reading AI's detailed and accurate feedback on that exam paper is not the same.

Like reading a commentary of a movie not being the same as watching it.

Like reading a summary of a book not being the same as reading the book itself.

When a teacher marks their students' exam papers, they have the opportunity to explore the folds of their brain.

A student's script. He was able to catch his own mistakes twice, but failed to detect the simple mistake while expanding brackets.

Knowing their background, a teacher's feedback may slightly differ.

To a student who makes frequent mistakes on expanding brackets, the teacher may emphasise this, and even consider a mini-teaching session or an additional exercise set for the student.

If this is not the case, a simple warning will do the job.

Can AI mimic this?

Maybe, if we feed it with all the exam papers of this student for the previous years.

But we don't "get to know" a student just from their exam papers, right?

We have a relationship in and outside of class.

We know more than just "this student frequently expands brackets incorrectly". We know the reason. We know the student.

Or do we?