May 23, 2024

How GenAI Might Support the Evolving Role of the Teacher in a Personalized Classroom

I was fresh out of the Peace Corps, having served as an education volunteer in Madagascar, when I took my first teaching interview at a middle school in Brooklyn, New York. I was navigating reverse culture shock and an abundance of purposeful energy as I prepared to start my professional teaching career. The principal looked at me and looked at my resume, sizing me up. He put the paper aside and said, “This looks good, kid, but let me ask you something: what are you going to do the first time a kid tells you to go…yourself?”
​
It was a different time. And it reminds me how long I’ve been in this profession. But what I remember even more than that question was what he said just afterwards: “Let me tell you something I tell all the new people: There’s no panacea in education.”

I didn’t understand what he meant at that moment. I was fresh out of university and further enlightened by my time abroad. I was a master of content and pretty sure that I was the panacea that education needed, the embodiment of best practice. I was John Keating meets Jaime Escalante meets Mark Thackeray, the sage on the stage who was going to save ‘em all with his voice and a piece of chalk.

I learned a lot that year. I learned that what worked second period didn’t work in seventh. That student X could have a great day on Monday, but be a disaster on Thursday. That student Y was awesome when we did creative writing, but would completely shut down when we reviewed grammar. And I learned that my principal was right, that despite what my resume said, I didn’t know what to do when student Z told me to…

Thankfully, my arrogance tempered a bit. I visited classrooms. I learned to ask questions. I saw a lot of what worked, and I saw a lot of what didn’t work. At that time, we didn’t really talk about impacts. Towards the end of my time there, “value added” had crept in via the No Child Left Behind Act, which was measured solely on performance gains on state exams from one year to the next. Day to day, though, we talked about what “worked,” and that meant that your class was safe and orderly, that most of your students liked being in your class, and that they performed better than average on the state exams. I managed all three parts, so I was a good teacher.

I was entertaining, and I cared, and I did a couple of things differently. For the most part, though, I chalked and talked, weaving my way through progressively-minded desk clusters. Kids had their own independent reading books that they chose from a classroom library, and they wrote their own creative pieces from a list of options I gave them, and we did an open mic on Fridays where they could share what they’d written.

The Open Mic evolved over time, and we incorporated music and stand-up pieces, and we had a rotating MC, and the students took the initiative to invite their peers and other teachers while I mostly sat on the sidelines. Looking back, that was the best part of the week, and I don’t know why it never dawned on me that more of the week could be like that. We aimed where we aimed, and that was as far as we reached.

Thinking back at those early classes, 30-40 kids in each, different capabilities, I think about how I would design that course today. Despite being a “good” teacher who did what “worked,” kids slipped through the cracks as I patted myself on the back for having above average test scores. In focusing on the aggregate, offering a solid median product, I missed opportunities with individual students. I didn’t realize that until I saw how impactful personalized learning experiences could be. When you aim differently, the places you reach can surprise you.



​Defining Terms:
Because there have been simultaneous movements to address this common concern, the terms can overlap in places and are sometimes misapplied, so I’l outline them here to avoid confusion.
Individualized Learning (IL)
Students progress at their own pace with a set curriculum towards common objectives.
Differentiated Learning (DL)
Students progress at their own pace with a differentiated curriculum towards differentiated objectives based on their needs and abilities.
Personalized Learning (PL)
A paradigm shift where the learning experience of each student is central to the creation of objectives and curriculum.
Universal Design for Learning (UDL)
A framework that guides the creation of personalized learning which ensures that all students can access, engage, and express.
Competency-Based Learning (CBL)
An approach where students work towards mastery of competencies, which are defined as knowledge, skills, abilities, attitudes, and behaviors.

As Audrey Watters, “The Cassandra of the Edtech Movement,” points out, personalized learning has been with us for much longer than the phrase has been popular. My students’ open mic was personalized learning before I knew it by name. Every Friday, one day out of five, my students were at the center of what we were doing. For the other four, it was content and curriculum in preparation for the state exam. It “worked”, but in hindsight the balance feels off.

Impactful Learning and How GenAI Might Assist:
In later articles I’ll go into greater detail of what personalized learning looked like in three very different schools in three very different settings. For now I’ll offer that while it can be a bit frustrating having to recreate the wheel in different environments, it’s also a meaningful challenge to try to get these balances right within different contexts, such that the needs of each student are met. It also keeps me on my toes, keeps me collaborative, and keeps me creative, constantly seeking ways in which to be more impactful.

Shifting from what “worked” back in my Brooklyn days to what I have evidence of being impactful today, John Hattie’s research on feedback and reflection stands out. Feedback takes a lot of time, though, and students need help with reflection. General use of GenAI tools and more focussed usage of GPTs can help with both, but they need to be incorporated thoughtfully, enhancing what we already do, not replacing it.

My first GPT, “Virtual Mr. Pultz,” continues to offer promising returns. My students like the immediate feedback. Students who don’t regularly seek me out in class or during office hours have reported feeling comfortable interacting with the tool at home. I do not plan on giving up written or face-to-face feedback anytime soon, but so far this tool seems to be improving the access that my students have to more immediate feedback.

To assist students with reflection, I built another GPT that incorporates our competency rubrics, our Student Directed Learning Contract, our Student Directed Learning Time Planner, and our Habit Intervention Form. The tool asks students which competency strand they would like help with. It then asks the student to identify specific instances where the issue presents itself. Once both parties seem to identify what cues the unproductive habit, the tool then takes the student through a habit loop reflection. From there the student is invited to create a new SMART goal for their SDL contract and to share their SDL plan for working towards that goal. I still want eyes on this process, but, again, these conversations increase access to a reflection “partner.”

In addition to feedback and reflection, advances in brain-based learning have shed light on some of the nuances of student engagement and how that can be impactful. While I’m sure it won’t surprise educators to learn that engaged students perform better, they might be interested to learn that students are most engaged when they are emotionally connected to a task or when they are experiencing something novel.

With this in mind, something that I’ve begun experimenting a bit more with is the construction of simulations. Ethan Mollick and Lilach Mollick put out a great article that showcases how they have built effective prompts. I’ve had my students work with their negotiating tool, the prompt of which is graciously shared in the article. I’ve also modified that prompt to be a cross-cultural scenario builder. In the former, my students gain a novel way to practice their English speaking skills while making cross-curricular connections. In the latter, they get to simulate a real-world encounter that challenges them to consider alternative perspectives, which, depending on what they input, can often lead to an emotionally charged experience. At the end of each interaction, students again receive feedback on their performance, and they are invited to try another simulation.

While I’ve been excited to experiment with some of these tools, I have been a bit frustrated with the uneven rollout towards access of frontier GenAI models. As this has been slower in some parts of the world, many of my students still do not yet have access to the strongest version of this tool, and some of those who do have access, have reported losing it after only a few interactions. To mitigate this, I’ve shared the GPT prompts with students so that they can use them in the 3.5 model. I’ve also tried gamifying the quick loss of access by making a contest of who can prompt for the best feedback within a certain amount of interactions. Hopefully this access will improve over time, but it is an area to keep in mind when offering these tools to our students.

Final Thoughts:
Like many others have stated, I do feel that GenAI will be impactful on education, but not in the way that some edtech companies are touting. Instead, I feel GenAI is best suited towards supporting personalized learning within existing frameworks such as UDL. In the case of UDL, the aim is to design experiences that increase student access to learning while enhancing both their engagement and expression. Adding a competency-based approach to that and supporting these initiatives with thoughtful GenAI usage, and we might be onto something.
In my own practice, first and foremost, I am and always will be seeking to build relationships with each of my students through a thoughtfully constructed curriculum that balances their own curiosity and passion for learning with the expectations of the particular context I’m working within. It’s a balance, and it’s one I’m constantly tinkering with, and I just got a shiny new set of tools to play with.
​
But any time I get a little too excited, I think back to that first interview twenty-four years ago, and I’m reminded of my principal’s words: “There’s no panacea in education.” He was right then, and he’s right today, but that doesn’t mean we shouldn’t aim a bit higher.
Next, week we’ll explore what it means to take a leap of faith with personalized learning—moving from theoretical potential to practical application.

---