Why Flint
|
Case Study
How Brophy Prep democratized creative and personalized AI for the classroom
Lulu Gao | lulu@flintk12.com
Mica Mulloy | LinkedIn
Mar 13, 2024
Case Study Summary
An introduction of Mica Mulloy
Can you introduce yourself and your role? What are you passionate about in the realm of AI?
My name is Mica Mulloy. I'm the Assistant Principal for Instruction and Innovation at Brophy College Preparatory, a Catholic Jesuit all-boys high school in Phoenix, Arizona. This is my 18th year at Brophy, and I'm passionate about keeping the pedagogical ball moving forward—making sure that our students are learning and that we are teaching in the best possible ways. Those ways may be, and sometimes very likely are, different from how we've taught before. This is not to say that tradition isn't important, and there aren't tried and true practices, but I'm passionate about helping students learn and have the best experience possible and then make the biggest difference in the world once they leave here. That has always been my practice, so when AI came about it was right up my alley here, and I dug in right away.
Brophy’s initial considerations about AI
When did you start thinking about AI?
I'm proud that we have been interested and aware of AI from the very beginning. At the very start of last school year (2022-2023), toward the end of the summer, we started hearing rumblings about some AI developments that were coming out, starting with some visual tools for imagery and then things like ChatGPT. I thought it sounded wild, and I'd believe it when I see it. Sure enough, in November of that year, ChatGPT dropped. There were a handful of us on campus who quickly said, “Whoa, this is game-changing and we've got to dig into it.” So we all started playing with it and talking about it a lot and trying out all sorts of things with it.
We pretty quickly got the whole faculty together to talk about it, knowing that it was going to have a big impact on schools in general. Our school did some demos of live, pre-recorded walkthroughs of ChatGPT. The sentiment was that AI was going to be game-changing for us as a school and potentially us as a society. We need to be aware of this and other tools coming out.
What was the sentiment like around that time, and how did you process all of those thoughts?
We watched in those early days as schools around the country started coming out with really quick, draconian policies of: “We're blocking this. Our students will never use this. This isn't going to impact us.” At Brophy, our thought was a couple of things. One, that's just not realistic. Sure, you can block it on your school network, but then what happens when the student goes home or uses the cell phone that all of them have in their pocket? And, two, do we want to block this or is this something that we should be leaning into and not only actively using ourselves, but also teaching our kids? How to actively use it and then how to not use it? What are the ways to use this as a tool for good and not as a distraction or way to cheat?
Eventually, we sent a statement out to the whole community acknowledging that AI is a thing and that we're actively exploring it. Kids obviously have access to it, and we just reminded them about our academic ethics policies as they are. Even with our approach to AI, when you turn something in with your name on it, you're taking credit for that work. So if you turn something in with your name on it and it's not your work, that's academic dishonesty. So we reminded students of that and then just continued to explore.
We put a faculty committee together to actively dig into AI use. By the end of last year, we published our first version of our AI philosophy statement on our website and shared it with the whole community. Essentially it concluded that we feel AI is something that we need to lean into. Jesuit schools and Brophy have a tradition of doing our best work on the frontiers. We think both the literal and metaphorical frontiers and the same goes for educational technology—that there's a lot that we can do really well as long as we're staying on the cutting edge of things.
And that doesn't mean just using technology for the sake of using the new, shiniest, fanciest spinning toy, but using it to teach our kids to make a difference and do something that they couldn't otherwise do or communicate in ways that they otherwise couldn't communicate. So we didn't have to change anything about our philosophy, and I feel fortunate to work somewhere where that's always been our way of proceeding with technology. In that sense, AI fits right into our school culture and that's our AI philosophy, that we're leaning into it. So, we put out that philosophy statement, and we made some revisions to our academic ethics policy to specifically talk about AI.
What are the highlights of the AI policy your school finally agreed upon?
Essentially, the policy is that students should not use AI unless their teacher specifically tells them that they can and that students can always ask. So, if a student has an idea of how he can use AI and the teacher has not specified that he can use it, then he just needs to ask. Teachers should then be receptive to that. It doesn't mean they have to immediately say yes, but it means that we're aware that kids are really good and smart and creative with this sort of stuff, and that kids are going to find productive, helpful ways to use AI that we haven't even thought about yet. We want to make sure that we're giving them an opportunity to do that.
We've done a couple of different professional development meetings with our faculty this semester, really digging into AI, where we had everybody explore AI tools. The idea is that we're not going to require everyone to use AI in the classroom or in their personal lives if they don't want to, but we are going to require everyone to at least be familiar with AI.
Our AI philosophy statement is going to have to be a living, breathing thing as AI continues to evolve. I don't think we can have honest conversations about AI until everyone has at least some experience with the tools that we're talking about. Otherwise, it's just all secondhand knowledge of well, I heard it did this, or I heard somebody did this. So we want everybody to have had some experience figuring out what it does.
I love the student respect piece you mentioned where you're treating them as a collaborative partner—how you said if the student asks and gets permission then they can use AI—because that really shows how open the dialogue is.
Yeah! I have a student committee of 80+ students who volunteered to be part of an AI committee. Those conversations are consistently some of the best AI conversations I have. What I told them was that I can't imagine us as a school making significant policy and practice decisions on a tool like this without actively talking to students. Because it's a tool that’s for school, it's about the kids, right? These are decisions that are going to actively involve them every single day now and in the future. So, gosh, we would be negligent, I think, to not involve them and make sure that their voice is an active part of these conversations.
What was Brophy looking for in an AI solution?
After you had a working policy on AI, what did your team outline as the key features or abilities of an AI provider that you were looking for?
I don't think we had a shopping list. We were just looking for what felt like it was going to serve our community best. I wouldn't go buy a car that way, but AI is such a new and rapidly evolving thing that I don't know what we could have written down. So, we looked all over the place and we tried all sorts of tools. I think I took more meetings with vendors in the last year than ever before because anybody who had a suggestion, I'd take a look at it. That's how I got connected with Sohan and Flint. Flint was from a recommendation from somebody that I know at a different school. I took a look and here we are.
That's good to hear that you took a survey of the AI environment. There are always new tools popping up, but you have to ask: which direction is it that our school should go?
We were looking for something that we couldn't just get on our own. Everybody could get access to GPT-3.5, so we weren't going to pay for a system that just used GPT-3.5. It's nice to use it in an enclosed environment, for sure, but I think what pushed us over the edge into making an investment is when GPT-4 became available. Those of us who were early adopters started playing with it and saw the significant differences between 3.5 and 4, and we wanted to be able to take advantage of the improved version. One, just because of everything that 4 can do that 3.5 can't, and two, so that the playing field for students was level.
We’re cautious with the use of technology on campus and try to have a level playing field so that socioeconomic differences don't show up through devices and tools that students have or don't have. And we became aware quickly, once 4 became available, that if we don't find a way to give everybody the same access, then students with means are going to have a much more powerful tool and therefore an advantage over students without means. And that's just not how we operate.
You answered a question that I didn't even know I had about why you decided to choose a school-wide AI platform versus just using a lot of various AI tools. There are a lot of free tools out there, so why invest in just one?
Yes, the equity was a really big reason. Once that was in place, the ability to create assignments within Flint and then structure those within an interface similar to a light version of a learning management system became really appealing. We got trial access and sent it out to some teachers around campus. The feedback was immediately positive with teachers saying: “Oh my gosh, I want to use this in class tomorrow. Can we add my students in so that we can try it out tomorrow?”
Why did Brophy choose Flint?
What other reasons prompted you to choose to use Flint, specifically?
Flint had access to GPT-4, and we saw the assignment generation tool as powerful and really easy to use. Technology can be intimidating to folks, and especially this massive, new, powerful technology that is just a flashing cursor in a text box and you're like: “What do I do with this? How is this going to write an essay for me? How’s this thing going to grade my students' papers? It's just a cursor.” The process of building an assignment in Flint was straightforward. When we started our Flint trial, that was the most explosive two-week period on campus in terms of AI use and interest that I had seen the whole time. People were anxious to dig into it and use Flint with their students.
Having shopped around for so many different tools, what made Flint click for you?
It was that both teachers and kids were immediately seeing the benefit. Again, we always try to ask students their opinions. I tried it out in the class that I teach and had other people try it out in the classes that they teach and talk to students about it. And kids said, “This is fantastic!”
We learned early on that some of them would need some training on how to talk to a machine, but that kids immediately saw the value in Flint helping them understand something or helping them learn or walking them through a process. And, it felt like, in a lot of ways, talking to a knowledgeable teacher, but they're all doing it at once and there's only one teacher in the room. Meaningful conversations with kids are fantastic, but it would take a long time to have those meaningful conversations with all 30 kids in your room. They can all do it at once in Flint. Also, teachers and kids saw the benefit right away that they had access to GPT-4. And the rest was: “Yeah, let's jump in!”
Once you decided that Flint was the way that you all wanted to go, how did you roll out the platform to the entire school?
The timing of it worked really well in that we had a built-in professional development meeting right at the time of the trial, so I was able to do a live demo for everybody. We had everybody in front of us all at the same time, and everybody was instantly impressed with it. So, we decided to jump in for at least a couple of months to try it out. And I said to teachers, let me know if you want to be a part of the Flint trial, and I'll add you and your kids in. Within the first week, I had added 800 students because people were so anxious to jump into it. So that was the roll-out.
Afterwards, it was just a lot of word of mouth, of a teacher saying: “Hey, I just met with Jim in the theology department, and he really recommended it. I saw him do this. Can I do that?” or “This English class is using it for this. I'd really like to try that. Can I do that? Can you help me with my first assignment?” So, it debuted at a meeting and then rolled out and then there was positive word of mouth. I think if I look at our stats now, I think most kids on campus are using it in at least one of their classes now.
How has Flint impacted teaching and learning at Brophy?
Now that a lot of teachers and students have been using it, what do they think about it?
Using Flint continues to be a really positive experience. We started a document among the faculty including Flint assignment prompts that yielded good results and must-have Flint AI rules that you should put in place. If you're looking for this, tell it to do this…if you experience this, try this…that's been really helpful for us to just actively share that and try it with each other. I'm impressed and proud of the willingness of people on campus to jump into this and learn through iterations of failure—something doesn't go well and we go back to the drawing board and then try it again.
I had a student come up to me a week or two ago, and it was a student who I would have never predicted to sing the praises of a technology tool. He said, “Hey, I have to tell you. Flint? Thumbs up.” I'm like: “Oh, awesome! Tell me more.” And he said, “We're using it in physics, and it's so helpful to understand the concepts. The way that the teacher has it set up, it's just been awesome. It's helping me learn.” Amazing. I was so thankful for that feedback. I've had a lot of conversations with kids and with teachers about that with similar experiences.
It's a brand new system, and so we've, of course, run into things where, like, oh, it would be nice if it did this, or, oh, it would be nice to do that. I think Sohan might be getting tired of our emails letting him know all those things, but the Flint team has been super receptive to any of those questions or pieces of feedback that we've had.
Can you speak to any sort of concrete benefit that you've seen from using Flint?
Long-term we don't know yet because it's still a relatively new tool. I have a lot of positive—albeit early and anecdotal—evidence where kids tell me the really creative ways that they're using even just the chat feature.
A student just sent me an email yesterday, saying: “I want to let you know that Flint did a very good job confirming my science fair results by analyzing the tables and graphs that I provided, which is something I did not know AI could do.” There are things like that, where students put their data in and get confirmation what they did right or wrong.
Another student told me that he was doing test corrections and he got his test back and didn't know why he got a question wrong. He didn't understand how to do it correctly. So he just put it in the chat, wrote here's the question, here was my answer, where did I go wrong? It told him the mistake he made and the correct way to proceed. I'm asked him, “You understand it now?” He's said: “Oh, yeah, easy.”
Or the number of students who said, I didn't understand this thing in my chemistry class. So I asked the chat to explain it to me in a different way, and it did, and now I understand it and I moved on.
The number of times that I've heard positive feedback like that from the student side of things is, I think, significant because AI's changing how kids can access information. Everybody has access to the Internet already. So, yes, the answer is probably out there somewhere. But there's a lot of information out there, so for a student to have to sift through that and realize, “I just watched a 15-minute video and I don't even know if that's the right way to do it yet, it still didn't help me understand this, it's too late to email my teacher at this point, I can't get in to see her for office hours until Thursday, and I don't know what to do”. That student has a tutor now in many ways.
The ability for kids to have a conversation that they, up until now, could only have with a teacher all at the same time is helping them progress in class more quickly and frees up the teacher to have more meaningful conversations with the kids. It's not replacing those meaningful conversations. If anything, Flint's creating time for that teacher to have better interactions and a better, deeper relationship with those kids because they're not caught up having to explain the same thing over and over and over again. So, lots of benefits like that that I've seen so far.
Final thoughts and the future of Flint
Is there anything else you would like to add?
We're really thrilled with Flint. I would say we anticipate using it in the future. The caveat there is that I don't know what's going to exist in two months, but I think our approach and our philosophy is going to stay the same and that we're going to do our best to make sure that we have access to the best tools to support our students. Sohan mentioned the other day that Flint's not tied to one specific AI provider, right? So right now you're powered by GPT-4, but if Gemini in the next couple of months surpasses that or something else pops up, that is even better that you're open to moving to that. I think that's reassuring, too, knowing that the tool that is working really well is always going to be fed by the best possible engine.It's the best-of-everything sort of approach.
What to read next:
How Merion Mercy provided students with trustworthy AI writing feedback through Flint
A case study of Flint’s partnership with Merion Mercy Academy in Pennsylvania, a school that has fully leaned into using AI to give feedback on student writing.
Human Connection in an AI-Enhanced World: Through the Lens of Maslow’s
What does self-actualization look like for a student's education? How can educators use AI to emphasize human connection?
What Schools Are Asking About AI — Part 2 (Eric Hudson)
Responses to common questions anout AI that Eric Hudson has come across in his consulting work with independent schools.
