What Students Gain When Teachers — Not AI — Grade Students’ Work

This article is part of the collection: Teaching Tech: Navigating Learning and AI in the Industrial Revolution.

During our research project on teaching and learning with AI, Mi Aniefuna talked to a lawyer-turned-teacher-turned AI ethicist. Masheika Allgood, founder of AllAI Consulting (pronounced “ally”), shared a story with me about her most transformative year as a teacher. What she did to help her seventh grade ELA students is something that generative AI, as we know it, can’t do. As the number of teachers using AI for tasks like grading and lesson planning increases, Allgood advocates that they be informed users.

All first-person accounts in this article belong to Masheika Allgood, with research support from EdSurge researcher Mi Aniefuna.

I’ve always taught. In undergrad, I was an elementary school substitute teacher. In law school, I volunteered at a preschool. In 2009, I was an online professor at Strayer University, when online learning had just become a thing. Most recently, I taught a course for Executive MBA students, and I am currently teaching a course for Juris Master’s students.

So, that’s what I do — I teach, and it’s what I’ve loved since I was a young educator. Among all my teaching experiences, my most formative period as a teacher, when I developed a style and pedagogy, was the year I taught seventh-grade language arts at a public middle school in South Florida.

By the time I stepped into that middle school classroom, I’d already completed three of my four degrees. My goal wasn’t to just make it to the end of the year; it was to help each of my students come to love the classroom as much as I did. For me, that journey began with preparation, ensuring that every student had a strong foundation so that when it was time to fly, we could all rise together.

As with any journey prep, this one started with taking inventory; in our case, it was diagnostic tests. I know from experience that smart kids can fudge their way through skills they haven’t fully developed, and education doesn’t always notice. I also know that people often make incorrect assumptions about low achievers, namely that they’re equally low-performing in all areas.

But people can surprise you, especially children. And you can’t properly assess inventory if you don’t actually check the cabinets to see where things are. For diagnostic testing, I selected specific areas to assess based on the year’s learning goals and the fundamentals students needed to meet them. For example, a student can’t analyze a section of reading if they don’t know how to compare and contrast. So I assessed, gave feedback and coached my students to incorporate it.

Then, I meticulously analyzed the assessments, comparing them across all my classes, and created a mapping system to visualize where all my students were. I even used the assessments to develop my lesson plans. And while I taught the same fundamentals to every single class, I would lean in heavily on a particular section depending on where the students were stronger or weaker.

After four or five weeks, we reassessed those foundational skills through a variety of means: they composed a song to explain a grammatical point, created crossword puzzles on key points in their reading and identified audio foreshadowing in movie clips. I also administered quizzes and assignments. By the time Christmas break began, all of my students had mastered seventh grade fundamentals. And in the second half of the year, we were finally able to fly.

Students read a fiction book from cover to cover as a class. For the vast majority of my students, that was the first time they’d done so. They enjoyed it because we did cool things that resonated with them. They drew timelines of important events we discussed, and each class developed an official timeline. They wrote letters as older versions of the characters, offering advice or wisdom as the book versions of themselves. We also held an official debate in which each class addressed the legality of mailbox baseball. One of my students researched the federal code on mailbox tampering and cited it in the debate, and that was one of the proudest moments of my teaching career.

And it wasn’t just my honors classes that flew. I also had students with learning and behavioral disabilities, hearing and speech difficulties and a few students who’d spent an inordinate amount of the school year sitting in the principal’s office. Still, they came to my class, and we all flew — because we all could. Because they spent the first half of the year learning the fundamentals, they were able to handle the higher-order concepts, having developed the grammar, vocabulary, and critical thinking skills that were necessary. Thanks to the diagnostic tests, I created careful and intentional lesson planning and reassessments that helped me close the gap.

During my observation at the end of the school year, the observing teacher couldn’t tell the difference between my honors and other classes. She said they were all performing at an advanced level. One student, in particular, who was such a regular at the principal’s office that they were shocked when I called in and demanded he be allowed to attend my class, not only passed my class, but he and every other student I taught also passed the state exam.

That’s why this research series matters. I’m passionate about diagnostics and assessments because I’ve seen what they can do. Because every student can succeed if they learn the fundamentals.

For me, the question that centers my teaching practice is: How do we know that students are learning? That question also drove my participation in this research project. In an era of nearly ubiquitous student learning platforms, such as learning management systems and other educational technology tools, how are diagnostics and assessments conducted? How do they inform the lesson plan and course goals? How is education changing, and are these changes improving the learning process for students?

In a district where students have been counted out and labeled as “low achievers,” Allgood utilized her expertise as a teacher, along with research-backed pedagogical strategies, to personalize instruction, deliver targeted feedback, make content culturally relevant and lead with empathy.

Fast forward to today, schools have deployed dozens of edtech tools to support teaching, learning and assessment, but many edtech tools now have AI-powered features. With competing priorities and dwindling resources, teachers are using generative AI to assist with feedback. What happens when we rely on AI to assess student learning and grade their work?