Join Jocelyn Wright, our Director of Community and Market Development, for a conversation with Liz Badolato, Associate Learning Experience Designer, on how she uses AI in her design and data analysis work and why she insists on understanding what's happening inside the black box.
At Six Red Marbles (SRM), we believe in continuous learning. Each member of our Learning Experience Design (LXD) team has selected an area of specialization this year to deepen their expertise and expand the value they bring to our clients. Our LXD Spotlight series highlights what our team is exploring and how those insights translate directly into client impact.
Today, I'm talking to Liz Badolato, Associate Learning Experience Designer, about how she's been using AI. Over the past year, she has gone from a physicist who was once wary of AI to one of its most thoughtful users. She shares how she decides what to hand off to AI, what all of this might mean for students, how AI-assisted analysis helped our team manage a 350-course project, and where AI output still falls short.
Your specialization is applied data analysis, which isn't a typical skill set for a learning experience designer. What does it mean in our context, and what drew you to it?
In general, I think of data analysis as taking a large amount of data, which can be anything from numbers, like counts of things, to categories, and finding patterns that can help inform whether or not what you're doing is working.
I have a background in science, and I'm currently working on a PhD in physics. Surface level, it looks like there's no connection between that degree and learning experience design. LXD doesn't do too many things with science, but data analysis is a big part of both.
Do you use AI in your data analysis or other LXD work?
AI is funny. I never thought that I would, but I've become team AI. I use it a lot now. But when I was in academia, I saw it as people trying to replace things I was doing, which I think is a common fear with AI.
People are afraid. I was afraid of AI coming into this job. But the fact that AI can do things like data analysis, I think, is really cool and useful. It makes things go a lot faster. It makes things more accessible to people who might not have even thought they could do these types of things two years ago. But I want to make sure to have some knowledge of what it's doing behind that black box.
How do you decide how much to hand over to AI?
I'm kind of newer to Claude. I use ChatGPT a lot more. With Claude, it does the whole thing for you, which is beautiful in some sense. But what I've noticed a lot, in workbooks either I created or teammates created, is that there are often columns it chooses to include, and I don't really understand why. The reasoning behind why things are there and what they mean is a little bit lost.
So I prefer to write my own workbooks and then ask for help. As an example, I keep a spreadsheet of my due dates for all the things I'm working on, and I usually manually highlight things that are due in a week. It was getting annoying. So I went to ChatGPT and asked, "How can I auto highlight things that are due within a week?" And it tells you the command you can use. So next time I want to do it, I can do it without asking for help. Whereas with Claude, it would do it for me without explaining what it was doing, and I wouldn't be learning as much.
The downside of AI doing it for you is if you don't understand it, it's hard to make small changes even, because it has to redo the whole thing most of the time. And if someone wants you to explain how you did what you did, it's not just a black box. It's something you can actually articulate.
"If someone wants you to explain how you did what you did, it's not just a black box. It's something you can actually articulate."
Where would you recommend that someone who's interested in using AI to do their own data analysis start? Do they need to get a PhD in physics, too?
Definitely not. And I don't recommend doing that while you're working full-time, because it's a lot of work.
If you're really interested in data analysis, knowing how to code helps, and knowing math helps. Even when you're not explicitly using numbers, there's a reason they make computer science majors take advanced math, even though they're not usually explicitly using it. That type of logic is part of math.
So this reminds me of a lot of workforce conversations, where the assumption is often that students just need to know the tool (like Python). But those background skills, like the thought process behind math you mentioned, end up being the best way to learn how to think, because you can't abstractly teach someone how to think.
Yeah. And that goes back to the black box thing, where if you have AI write Python for you, if you don't understand why the code's doing what it's doing or how it's doing what it's doing, you're going to miss a piece of it. There'll be more work in the beginning for you to fully understand it, but it'll be less work as you continue to work on the project.
In science, I've noticed a lot of times people know what answer they want, so they miss things on the outskirts. Whereas with the more humanities, big-picture look of things, they're more likely to catch some of those outliers or be interested in the outliers. But for the analysis part, I think all different kinds of brains can be useful in that.
The humanities, and the writing question specifically, is where a lot of people worry about AI. The humanities person in me thinks at least three-quarters of writing is thinking. How do you see AI changing the way students learn to write?
I still don't know how I feel about students using AI so heavily. But being smart about using it is going to be how we write in the future. Not everyone, but that's probably how academic writing winds up going. I don't know that I love it, because it's not how I was trained and I prefer to write things myself. But I have a feeling eventually you're not going to be able to tell the difference between what students wrote and what AI wrote, and you're going to have to just accept it and change the way that you're analyzing or grading how they write. It's more about how well they used the tools available, not how well they wrote this totally on their own.
I don't know what I would do if I taught English again, because if you assign essays, a lot of students, most of them probably, are going to use AI. You're going to get 20 of the same exact essays, or nearly the same essays, if they all use AI, especially if they do the laziest thing and just copy and paste the prompt in. And a lot of schools don't allow you to use AI scanners because they're not accurate. At least in physics, you can give in-class tests, and they can't use AI when they're sitting in front of me. You can have in-class essays, but it's not the same. You can't do the planning or the research in class.
You used AI quite heavily to manage a recent client project. How did you find that balance between AI and human work?
This project was a little bit chaotic. We were accelerating courses,* which in theory should have been 16 weeks, but a lot of them weren't. The courses all kind of came in in a rush after months of planning. So there was a lot of scrambling, and I saw an opportunity where I thought that more structured data input would help.
I wound up taking over provisioning, which meant I was the first one to look at all the courses we were getting. I used AI to estimate the page counts and to compare the courses to determine alignment of content. So it was partially AI-driven data analysis and partially my eyes looking at things to get a rough tier. The tier was really meant to say how long it would take a typical instructional designer to build this course.
It did, I think, save our team lots of time. By the time the courses were really rolling in, we had a relatively smooth, streamlined process for that intake. We ended up doing 350ish courses in about 3 months.
This sounds like a really useful application of AI, but we all know AI isn't a miracle worker. Sometimes the output doesn't hold up. Can you share an example of when you've encountered this?
I've seen lots of AI-driven documents where sometimes I'm like, "This doesn't make sense." And it's not even necessarily data. Sometimes it's just a Word doc with instructions, and I'm like, "These instructions are not clear." Like step-by-step directions where it'll say, "Pick this," and that's not an option. So I'm using my judgment to figure out what's the closest thing to that. I can usually figure it out. So it's saving the person who wrote the directions time, but then it's taking me longer to make sense of them.
So I'm still not fully convinced that Claude is a replacement for producing your own documents. It's faster, and I get the benefit of it, especially if you're handing off internally and it's someone that can read it critically and not just take it at face value.
Reading critically depends on having the skills to begin with, which brings me back to our conversation on students. I saw that the University of Chicago has banned AI in its social sciences core curriculum. They're not necessarily anti-AI; they have a partnership with Anthropic. But their stance is that when you're building these core skills at the beginning of your college career, you need to know how to do them without AI. It reminds me of high school calculus. We did something the hard way for three weeks, and then the teacher was like, "Oh, and here's the formula for how you can do it in 30 seconds." I still remember that sense of frustration: Why didn't you just show me the shortcut from the beginning? But it makes you appreciate the shortcut, and you understand how it works.
Yeah, it's good to understand what's going on in the black box before you use the shortcut.
Thank you so much for chatting with me, Liz!
What Liz's Insights Mean for Our Partners
The goal isn't to hand everything to AI. It's to know what AI can do well, what needs a closer look, and where human judgment still matters.
That's AI + Human Insight.
Explore SRM's AI + HI Hub →
Liz Badolato
Associate Learning Experience Designer, Six Red Marbles
Liz Badolato has an interdisciplinary background spanning physics and English. Before joining Six Red Marbles, she spent several years in higher education, teaching courses in physics, astronomy, and writing while developing hands-on labs and research-focused curricula. Liz's path to instructional design reflects her interest in connecting technical and human-centered ways of thinking. With a PhD in English and a second PhD in Physics in progress, her work bridges scientific problem-solving and clear communication. She brings that perspective to her projects at SRM, where she focuses on designing learning experiences that make complex ideas accessible, engaging, and meaningful for a wide range of learners.
*"Accelerating courses" means shifting a traditional 16-week format down to a shorter timeframe while maintaining rigor and student engagement. Learn more here.

