AI for Faster Learning Design With Dr. Dwayne Wood

Welcome & Series Context

Jackie Pelegrin 0:00

Hello, and welcome to the Designing with Love Podcast. I am your host, Jackie Pellegrin, where my goal is to bring you information, tips, and tricks as an instructional designer. Hello, instructional designers and educators. Welcome to episode 156 of the Designing with Love Podcast. I'm thrilled to welcome Dr. Dwayne Wood back for part four of our mini-series. In the first three parts of this series, we explore generative AI's impact on academic curriculum, multimodal strategies for adult learners, and the adult learning principles that should guide effective education. Today we're bringing those ideas together as we discuss how to embrace AI as a tool for innovative instructional designs. This conversation is not just about using AI because it's new or popular, it is about thinking carefully, creatively, and responsibly about how AI can support better learning design. Welcome back to the show, Duane.

SPEAKER_01 0:58

No, glad to be back. Thank you.

Jackie Pelegrin 1:01

Yes. I almost stumbled over my words there, and I was like, oh, well, that's the that's the authenticity right there. People know it's not AI when I make a little mistake here or there, right? Yes.

Rapid Prototyping For Faster Iteration

Jackie Pelegrin 1:16

So to start, uh when you think about AI as a tool for innovative instructional design, what possibilities excite you the most?

SPEAKER_01 1:25

Well, so coming from um you know my experience in instructional design, one of the things I really like about, or what what I think this um brings capability, this technology capability, is the idea or the the ability to prototype very quickly. Um so that you you can increase the speed of your cycles, you know, and produce some prototypes uh much quicker than you would be able to, and actually have maybe a wider range of options to think through for your particular project that you're working on. You know, I I I've learned over time, you know, working with a client or you know, uh on a project uh or you know, some kind of learning event or learning uh course. And I always find that the client sometimes really doesn't know what they want, um, but they know what they don't want. So, you know, that that rapid prototyping process worked well there because you could do something really quickly that you, oh, that doesn't work. Okay, all right, let's go to the next one. And you didn't spend a lot of time and effort to do that. Well, now with this technology, you can do that even quicker. Um, you can you can run a whole bunch of options. And it and it, you know, the the the value is not that the you know the AI is gonna produce a perfect example of the first couple iterations. What the value is, is that it gives, you know, uh, like me as the designer, a bunch of different options that I might not even have thought about um right up front and real quickly to start thinking through that project and how I'm gonna achieve like maybe I'm gonna combine some of the ideas that it offered, right? Um, you know, maybe I'm gonna apply them a little bit differently than it was given, but it sparked that brainstorming and then ideas as we go and um as I start building out um you know that learning event. So I really think in that space, um, this technology is is is gonna be um super powerful for a designer. Um it just speeds up that um creativity space, right? When you're when you're sitting down that initial design phase and you're thinking through those things, you can again wrap a prototype really quickly.

Jackie Pelegrin 3:18

Right. Have you noticed that that's uh been able to cut down on the development time a little bit when they're able to utilize these tools for for different types of uh creation and and iteration? Have you noticed that?

SPEAKER_01 3:33

So initially, I probably was spending the same amount of time. And and I I tell you that because I was I was still learning the tool. I was learning how to ask and get what I was looking for. Because um, because what was happening initially is I would ask the question, I'd get a response like, that's not even close. So then I'd I'd prompt something else, and you get this like long chain of prompt, and then all of a no all of a sudden you're like one hour into it, you don't have anything, right? Um but I started to I started to develop and like know how to do it and then start building myself templates, start building myself like um I call them assistants, right? So I can like run an assistant, it'll ask me some questions, I fill in the answer, and then I get a starting product. Uh, you know, where and it helps me with that process. So that whole beginning prompt stuff, I've I've I've I've sped up the productivity there. I don't know, I'm not worried about inventing it now. It's already invented, and I'm just using what I have now. And to so I I think we have to be careful when you're thinking about initially putting into a process, is you need to make sure that you, you know, um the user that's using the tool understands how to use it. You know, maybe you have some templates or stuff worked and you know, have thought through that process a little more than just injecting a tool into a specific spot in the process.

Jackie Pelegrin 4:48

Right. Yeah. So it has to be something that is natural and makes sense and it's not forced, right? Yeah.

SPEAKER_01 4:56

And and on top of that, right, it when you do this, you're not just accepting whatever it produces. You're putting a designer eye on it and thinking through it, understanding what your outcomes you're trying to achieve, right? And you're using that as an evaluative tool of is it meeting it? Am I changing it? Am I rejecting it? Right. Um, so there's there's that level that you have to put instead of just, you know, yep, produced it, good or no good, right? You got you got to put that um that evaluation and think through it.

Jackie Pelegrin 5:26

Right. Because the tool doesn't know the audience like we do. It doesn't know certain things, right? And certain nuances that we would know.

SPEAKER_01 5:35

Yes. You're absolutely right. You know, no one knows that audience better than you do, right?

Jackie Pelegrin 5:40

Right. Yep. We still need that human judgment. Yes.

Learning To Prompt Without Wasting Time

Jackie Pelegrin 5:44

So I wanted to talk a little bit about how AI can support the instructional design process in practical ways. So for you, Duane, how can instructional designers use AI during the design process, such as brainstorming, content development, scenario creation, assessment design, or even learner support?

SPEAKER_01 6:04

Yeah, so when I when I think about this, right, we kind of talked about that iterative process, but you know, if if we looked at, you know, ADI, um, I mean, AI can have a role in all of the different phases. Um, you know, we think about analysis. Well, if you have a bunch of data, it can help um analyze that data, right? It's very good at pattern recognition and and so you can actually use an analysis phase to better understand the audience. You know, maybe you have a bunch of like quiz results or test results. Again, you don't want to use personal information, make sure that all that stuff's gone, right? But you just have the data. Maybe you can help the system identify highly missed answers. So then that indicates where you need to maybe put some more emphasis on in the design, right? To meet that that objective, right? So there's there's some there. Um we could talk about um you know different ways, like again, the brainstorming. Hey, I have this, I'm I'm I'm doing this topic and I'm thinking like this. Um, can you uh suggest three different ways that that this could happen, right? And just again, that brainstorming um piece and see what it comes up with. Uh I I got one prompt and I'll share it with you, Jackie. That um I I asked, I I wrote it out to help me think through how do I make assessments AI resistant? Um it does produce some decent ideas, some some of them are just silly. You're like, okay, all right, that's just not gonna work. Um but some of them were like, oh, that's a that's a not a bad idea, right? And how do I apply that in my context, right? Um start thinking through that. Um what I like the idea about uh on the student side is you know, if if you can provide um these tools can provide some level of scaffolding, right? For uh a student, but if it's not, it needs to be um set up to where you know the student's still gonna be able to exercise that cognitive effort you are while they're using the tool and not just provide, you know, an answer, right? Um so I like to I I sometimes will build things, um I call them a Socratic engine, right? So it you know, if you put in a question, it's gonna give you another question, right? It's it's gonna probe and it kind of get you and uh get you going in the right way to to find that answer and to encourage that critical thinking um piece of it. You know, so by itself, if we just give a an open AI tool to a student for support, uh I I'm a I'm a little concerned there because it we we probably are gonna maybe reduce that cognitive effort completely, right? A student has the ability to kind of copy and paste the assignment instructions and hit go and then you know copy and paste into a a and and then the cognitive effort that you're trying to achieve, it just didn't happen, right? Um it's probably gonna produce a pretty decent polished product that might be hard to tell, right? That was actually produced by an AI tool. So um, but in specific ways um for student support, I definitely could see where it could be used um there. And then in the evaluation phase, again, you're collecting data. And these and these systems can help analyze that data and may pick up on things that we wouldn't have thought of um, you know, or or as we're you know looking at the data

Practical AI Uses Across ADDIE

SPEAKER_01 9:11

and what it means and what does it mean in our design and iteration.

Jackie Pelegrin 9:53

Oh, so true. I love that because we you know I do survey development during the needs assessment phase, and what I do is take those that's the output of the survey from my faculty and I'll stick it in our AI model and I'll just have it give, you know, have it look for those patterns and those key themes, and I'll use that for my program kickoff meetings. But it's really nice at the end of that too, when we can look and see, okay, now the students are in the courses, they've been in it for one or two semesters, right? And then we can look at all that that data from the faculty, from the students, from at their end-of-course surveys, uh, because we have faculty and students doing end-of-course surveys. So it's really helpful to see, you know, is this an issue just on the faculty side, or is it are the students having trouble and where are they getting stuck, right?

SPEAKER_01 10:41

So it's really helpful to if you if you can identify those pain points, right? Um that that's super powerful. Uh and right and and it can do it much quicker. Um, you know, you just have to sit down, sit down and think of, you know, what data am I, you know, make sure the data's clean and you know, and develop out what that analysis looks like and have that, you know. But yes, it could absolutely produce some really some really insightful um information.

Jackie Pelegrin 11:04

Right. Instead of you trying to, as a designer, dig through all that information, which is what we used to have to do. And I still like looking at the raw data, but when I want to look for those patterns and those themes, I I really like being able to utilize that during surveys and observations, things like that. And and also um just, you know, being able to wade through all that all that data a little bit better. Yeah.

SPEAKER_01 11:30

Yeah, absolutely. Yeah.

Jackie Pelegrin 11:32

Right. You know, and I I'm wondering too if it can be used for when you're when you get ready to implement something before uh, you know, while you're testing it with your learners, you can even um utilize some of it for that to see what came through, right, when you tested it out with your learners. What were some of the patterns that came about? What are some areas that that are low-hanging fruit, right, that I can adjust in the training before or in the course before it gets rolled out to the larger group of people.

SPEAKER_01 12:00

And you know, um, what I've done virtually, um, so if we have like a virtual synchronous session, is I've actually had a tool that I've set up where um we may be discussing something, right? Um and I have the tool set up that way. So we we actually everybody uses the chat, copy and paste into the tool or into the AI, and then we can have it summarized and and we can see that happen in real time um in our discussion online. And then we can use that and we build off that and you and you can actually produce a product at the end of so I I've I've done that as a live facilitation tool, and it's worked really well. Right. So I mean, there's there's a bunch of different ways that I think it could be used. Uh one of the things I like about the live facilitation tool is let's say your learning extends beyond that session, right? You share that with the group, and then they get to work with that for other pieces, right, of the um the learning event. Uh so again, there's a bunch of ways to kind of think through that where this could you could leverage this, you know, where we could do that in the past, but it would just be laborious, right? We'd have to, you know, copy and paste the transcripts and put it all into a format and where this does it live, right? And it produces right there on the screen. So uh again, there's there's a there's probably a bunch of ways I haven't even thought of yet, right? That that um we could be used to leverage this. But again, thinking in the way that it's not replacing, right? It's augmenting. And right I all the time when I get to be ready to use a tool, I ask myself, is this gonna replace what I'm what I'm thinking, or is this gonna help augment my thinking? And if I can't answer that the way, then I probably don't like to use the tool. Uh because you know, I want to make sure, right? I'm trying to make sure I'm using this in a very responsible way and in a way that it's not gonna erode my expertise. Um, I want to maintain that. I want to stay sharp. Um, you know, so I need to encounter problems and work through them. I don't need somebody to just give me an answer.

Jackie Pelegrin 13:56

Right. Absolutely. That's so true. I think this segues into our next uh part of the of what we're gonna talk about is um

Human Judgment Over Trust & Hype

Jackie Pelegrin 14:06

one of the key questions I think that comes up for many designers is how to use AI without losing the human judgment that good design requires, because we don't we don't want to lose that, right? As you're talking about. So for you, Duane, what does it look like to use AI as a creative partner while still keeping the instructional designers' expertise, judgment, and human insight at the center of that process?

SPEAKER_01 14:30

So what's really unique about this technology, and we've we've talked about it, I think, in a previous podcast, is that unlike other kinds of technologies, this one is right in the human cognition same area, right? So there's a real high risk of substitution, right? Just having to do it. So when I really I think we have to, we have to have a different mental model than we we have. And what we have, honestly, I I you have to treat it as a tool. If you start thinking about it, I think as a true partner or as like a colleague, or um, I think there's room there to now where you start to develop trust that you might not want there. Um, you also might think of a process where you're substituting that tool as a partner as part of your process when it it probably shouldn't be. It's a tool. So when I think about this, is you know, um, we can still think about it as forming a team, you know, and think about the idea of, you know, why do we form teams? It's because everybody has strengths and everybody has weaknesses, and we form teams so we can offset each other's weaknesses with each other's strengths. So, same way, same way here, you know, between human and machine, um, the machine can crunch all kinds of large amounts of data, right, pretty quickly that is hard for us to do. So, let's leverage that capability, right? Um, but it doesn't know what truth is, it can't exercise human judgment, right? So those things reside with the human. So we just have to kind of think through that and not really start thinking about this as like uh a colleague, right? Um it's a tool. Um and we have and just like any tool, we have to know how to use it. So that's where I think we really have to start thinking about these um or this technology, because if we start thinking about it as uh, you know, it's my colleague that's helping me out, I think we start attributing that um certain um things to that that end up hurting us in the long run, right? The person that's accountable for anything this produces is the user. It's not the tool, right? So um if we start attributing as a partner, we start attributing that the response we we take less responsibility of whatever it produces. And we don't want uh again, my my thought is I don't want that to happen. The user of the tool should have full accountability or has to have full accountability. Um so we have to make sure we are thinking about this tool in the right way right from the beginning. Otherwise, we'll we we I think we walk into maybe some some pitfalls that we have to be careful of.

Jackie Pelegrin 17:03

Right, exactly. Because then people start thinking it's gonna be like the Terminator movies, right? Where all of a sudden they're gonna take over the world. And yeah, I try not to think about that. I'm like, no, I don't I don't think it's gonna get to that point.

SPEAKER_01 17:16

Yeah, we're a long way off. Like I honestly, when you say artificial intelligence, I think that's kind of a misnomer there. Um I think statistics. That's not really intelligent, it's statistical probability based upon patterns, right? So um again, I mean it it produces some pretty decent results. And I do have like some concerns about when you start letting an autonomous system make its own decisions. Uh there I I think we're a long ways where we start allowing that to happen, at least in in my mind. Um you know, but yeah, that's that's where I I think we there's a little bit of concern there. Um, but we just have to develop those mental models and how we effectively collaborate with this technology. Um and I am using maybe I shouldn't use the term collaborate. I should just use the term use. You know, how do we apply this tool? Um is maybe the better way to say it so we don't start thinking of it as a person.

Jackie Pelegrin 18:16

That's true. Because it could, yeah, you can start to muddy the waters a little bit, right? When when you do when we start doing that, yeah, it's easy to think of this as an intelligent tool. And yeah, that's uh like you said, a a partner. Yeah.

SPEAKER_01 18:31

Yeah, there's there's a lot there's research out there that shows, right, that the trust level um kind of is like a an arch, like if you graft it. So, you know, no trust, you're super skeptical, right? You don't get as good as ru results. Um, and then if you're like too trusting, you don't get as good as results, right? But it's it's it's a little somewhere in the middle. Um, you know, so just being able to understand and and work with it as it's a tool. And I I think if you just take that pers that kind of perspective and and you start applying that to your processes, I think you'll see where you know you start using it in a way that um, you know, you're not gonna you're not gonna use something where it has a hallucination or it makes up something, you're gonna catch it um when it's doing that.

Jackie Pelegrin 19:12

Right. Absolutely. Wow. That's so important, Dwayne. I'm glad you brought that up. That's yeah, that's that's important to make those distinctions. And yeah, that's why guardrails are so important and and having rules around these types of things, it's not to limit us, but it's to help us understand what the limitations are of these tools and what they can and can't do. Yeah. Or what we want them to do and not to do.

SPEAKER_01 19:37

Yeah, and you know, and in the it's the age-old question of like, you know, just because we can do it, should we do it?

Jackie Pelegrin 19:43

Right, right. Yeah. Then you start getting into the ethical and and all of that, yeah. Those types of things. Yeah. Where it can it can get sticky and yeah, absolutely. Yeah. So another area where AI is often discussed is the personalization and learner support. I've heard this so much among my

Personalized Learning With Scaffolds

Jackie Pelegrin 20:02

team and with other designers. So how Duane, how can AI help designers create more personalized, adaptive, or practice-based learning experiences? You even mentioned problem-based in one of our last ones. So feel free to kind of talk about that too as well.

unknown 20:18

Yeah.

SPEAKER_01 20:18

So when we're looking, you know, one of the things we talk about in adult learning, right, is the idea of having authentic tasks, right, that are relevant. Um, so this tool can actually help do that, right? So you can take stuff and you can help develop some of those um scenarios. One of the things I've done with the tool that's been pretty successful is building branching scenarios where it, you know, before AI, building a branching scenario was a pretty heavy lift because you have to factor in all the different decisions and start breaking out the tree, right? And working all that, and then build that into a system, right? Into, you know, uh captivate or storyline or something. That's a pretty pretty significant lift. But I could build a branching scenario in minutes now that you know that you put in your feedback, it gives you a response, and it starts branching out, and then you can have it and analyze your decision making, right, as you're going through it. So I think there's there's a lot of potential there. And that ends up being very personalized to the user because they're inputting, it's reflecting off exactly what they're putting in. Where we have to be careful is um if the system is not developed where there's guardrails, like you said, it could go way outside of trying to achieve the outcome you're trying to achieve, that learning outcome. So we can't lose that focus on achieving that learning outcome. However, we're using the tool, that tool's being used to achieve that learning outcome. So we we we got to think about that. Another piece that I kind of think about this tool is if it's done right, you know, one of the things in, you know, is if you send somebody's gotta do uh work on their own where they don't have an expert in the room and um they're working through something and they're struggling, these intelligent tutor systems, right? Um you're gonna get immediate feedback when you need it. And not only is that I think gonna help prevent like you know, maybe developing bad habits or, you know, uh, but it's also I think gonna help with motivation, right? You're not you're not as frustrated. It's you're you you don't feel like you know you you can't do it. You're getting the the the support and the scaffold help when you need it in that time. If a because I I know in in my realm, right, uh in the asynchronous environment, if a student struggles, they send me an email. Well, I might not see it right away, even though I try to respond like within an hour. But then it might take time to set up a meeting where we can talk about and help them. So you I've already like extended that help timeline just because, you know, uh of availability. Um, where this tool, if it's done right, could be immediately available to aid that student. So I I think there's some potential there, but again, it's doing it the right way, right? And doing it in a way that still um allows that cognitive effort, right? And and augments achieving that outcome, right? And just doesn't replace right, absolutely. Which is nuanced, right? It's super nuanced and um it's a challenge. Uh and the answer is always I think it depends, right?

SPEAKER_02 23:15

What is all the context?

Jackie Pelegrin 23:20

That's so true, Dwayne. Absolutely. And do you do you think that some of these authoring tools like captivate storyline, do you think they need to have like those limitations built in where it says like um, you know, the tool should not do the or the branching scenario should not go this route. This is where we build in our guardrails, or um, where do you think that should kind of live for the designer so that they can have those proper guardrails in place or the do's and don'ts, right?

SPEAKER_01 23:47

Yeah. So I mean, you bring up a um, you know, something I've kind of been thinking about is because if I start using

Guardrails For Scenarios & LMS Chat

SPEAKER_01 23:54

an AI tool as an instructional tool, you know, who's building that tool? Is it the instructional designer? Which they're probably in the best place because they understand the outcomes. They're already working with the subject matter expert for validation. Right. So um, but yes, like so, like some of those like branching scenarios, I'll put in a bunch of restrictions. Um, a lot of times I'll put stuff like only use the content that's been uploaded. Right. Um so I just try to make sure that I'm narrowing what it will work within so I don't get something you know crazy um that's not related and it's still aligned with that objective.

SPEAKER_03 24:35

Right.

SPEAKER_01 24:35

Um, because I I actually I I I tend to usually write in a uh a kind of an end point where I have it analyze the the results of that branching scenario against the objectives. Um just to give some kind of like, Randy, you still need to look at it, right? You still need to look at it. You can't just accept what it produces, but it can give you an idea of um you know where where you're uh meeting those objectives or not. Um so I I think that's an unsolved question right now. Um, and maybe in and organizations that don't know where to go with this, is if we do start using AI as an instructional tool, who's putting the guardrails on there? Are are we are we just giving them to the student to do? Um, or is it gonna be an institutional like programming piece where it's like a closed AI system? Um, you know, so these are all like big questions, right, that we're trying to figure out. Um but I I know for me, if I use it as is AI as an instructional tool, I'm either gonna provide the prompt to the individual to copy and paste if I if I want them to use it that way, because then it has all the guardrails in it, right, for them to use. Um, or I'm gonna use like a live system where I'm controlling it, right? So then we can you know you get um again, that's just kind of my take on it right now when we when we think through these, but I think you're absolutely right. You know, what happens now if we're you know, um now we have LMSs that are have AI chat in them, and we're using that as an instructional tool and scaffold. You know, what role does the instructional designer have in how that tool is being used within that course? Because there's probably times in that course you don't want it to be used because you want the cognitive effort to happen. You know, so again, I this is this is breaking, right? We're still trying to figure this out. Um as as you know, educators are coming together and trying to figure out how do we make this happen. And we have, you know, we have policy guidelines, we have, you know, we have technology limitations. Um but you know, uh at some point we're gonna have to try to figure this out and what that means. Um and and again, what is the instructional designer role there?

Jackie Pelegrin 26:42

Right. Absolutely. Yeah.

SPEAKER_01 26:44

Maybe as part of your ID team now, you know, you got you got the ID, you got your media specialist, and now you have a machine learning AI guy, right? Or person that's on your team now. Because, you know, as you develop the what you want the tool to do, they're they're doing it for you.

Jackie Pelegrin 27:01

Right. That's that's true. Yeah, yeah, that might be the case. Yeah, I can see that being really helpful, you know, for training um for uh, you know, courses and things like that, um, building that in there. That's one thing we've we've been doing with these interactive personas, uh, where you know it's it's a uh where it acts like the client or the patient or something if it's for nursing. And then our media design team, they they actually have a like a template and they have the either the college representative or subject matter expert fill it out, and then they put, you know, what's the scenario look like? What's the the demographics of the individual? And then at the very bottom, they put like, what is this tool not supposed to say or do? You know, what are what are the limitations? You know, what are you know what do you want it not to, you know, you know, do or or what do you want it to do? Yeah, exactly. But that's one of the requirements is they have to put the limitations in there. Yeah.

SPEAKER_01 28:02

Yeah. Um because again, it'll it'll just go, right? If you don't tell it, it doesn't know.

SPEAKER_02 28:06

So absolutely.

Jackie Pelegrin 28:12

Yeah, that's so true. And even like you mentioned with the that's a you know, that's great how you mentioned like with the chatbots and LMS systems. Um the university that I teach for, um, they don't have that yet, but uh I'm sure it's gonna be coming. But that's true. I mean, there's certain things that you don't want that enabled for, right? That you want that cognitive effort because then, you know, users get if they get used to seeing it in there or learners get used to it, they're gonna think, oh, I can just go to this and click on that little bubble, right? And get the help. But sometimes you want them to work through that on their own. Yeah.

SPEAKER_01 28:47

And where I see it like really super important, or you know, when we think about this, is just when we talk about foundational domain knowledge. Um, because it it you have to establish that before you can actually evaluate a response from one of these systems, right? Because if you don't, if you don't have the domain knowledge, how do you know what it's telling you is not right or um without you know, you have to start doing some some deeper research. So you're more likely probably just to accept it. So there's already a risk, right, without having that domain knowledge. So you we definitely wanna we wanna think through, you know, what are the things where we want to use that tool where it's augmenting, um, but what are those times where we want to create that cognitive effort at the level we need to produce that domain knowledge?

Jackie Pelegrin 29:33

Absolutely. Wow, love that.

SPEAKER_01 29:36

And some of the things I think about is um, well, if we're thinking about that, maybe we need to think bigger in the idea of maybe our objectives now are not written in a way where it's a concrete skill or uh attribute, but it's a thinking skill. So we write them as something like, uh, you know, we're we're gonna do, you know, the course is on whatever, but there's a critical thinking objective. Uh, you know, so because we'll be using these technologies more and more, um now the student has to demonstrate that critical thinking um because that objective is there.

Jackie Pelegrin 30:16

Right. Oh wow. That's great, Duane. I love it. Wonderful. So I wanted to move into our bonus question. What is one AI-supported instructional design activity

A Simple AI Exercise To Try

Jackie Pelegrin 30:29

that listeners could experiment with this week?

SPEAKER_01 30:33

So I would tell them find one of their objectives from any courses they've ever done. Just pick one objective. Put it in one of these chat box and ask it for like um, you know, start with like, you know, how what what are three different types of assessments um for this objective? See what it produces. Um find one that when one comes up and it says, you know, what whatever it is, then ask, hey, I would like to explore that one deeper. See what it produces. And just kind of do go through that and see what it's producing, what it's giving you. Um, and then I think you can start, you know, expanding out on it, right? Um and start, but just get in there and try it. And um again, I I you really learn these tools when you get in and do it, right? You can take all the training all you want, but until you get in there and do it and actually start seeing how it goes and the flow runs, you just need to get in there. So that's what I would say. Take one of your objectives from any any course and just see what it can start producing, and then start thinking about how it potentially could become part of you know your um design process, you know, where it would be helpful or where it wouldn't be helpful. That's another good question to answer, right? When when not to use it. Um just get in there and try it. Uh, but try it with something concrete, right? Something you know so that you can evaluate against it, you know, and see what it's producing.

Jackie Pelegrin 31:51

Right. I like that. Yeah, something that you already know or that you're familiar with, right? So you can you can see how how well it's doing. One thing I I've done a lot with the curriculum is I if we're working on a revision for a course, I'll I'll take an existing uh discussion question or assignment. And if I need to tie it to standards, or if I'm like, you know, I don't think this fits the objective very well, I'll I'll stick it into our model and I'll give it the objective, right? Give it all the details and and just do the same thing, like what you're talking about. Can you give me three ideas to make this maybe more of an authentic type of assignment or discussion question? And does a pretty good job. I agree. And then you're right. I like that idea of honing in on one that you like and really digging deep into that. That's a great idea. Yeah.

SPEAKER_01 32:41

If you want like a way to extend when you do this, um, so if you if you put in like that one that one objective, right, and you're asking it to produce some three assessment ideas or whatever, before you hit enter, type in ask me questions before responding. And what the system do is it'll generate a bunch of questions, then just answer them. And you'll find you'll get a much more refined result. Um now, sometimes you might want to limit how many questions they ask you, uh, because you might get a hundred, right? So and you'll be spending the whole time trying to answer questions. Um I usually say, you know, answer, you know, uh ask me five, five questions. Or I'll say, you know, what what are what are five things you would like to know before responding? Right? Um and it'll and then it sometimes picks up on things that um you might not have put in the prompt. And you might get there, you might get to the the result you're looking for a little quicker.

Jackie Pelegrin 33:34

Right. That's true. I think that's important you brought that up with the the um putting those um, I wouldn't say limitations, but just parameters, right? Around give me five questions, because it's true. If I don't tell the tool, you know, make sure this DQ is um 150 to 200 words, it'll give me a mini assignment. And I'm like, oh, that's not what I wanted. It'll just go way too far. It's true. I mean, it it I don't know if it just the tool just wants to just spit out everything or just I don't know what it is, but I'm like, wow, that's way too much. So I've learned to put in those um those parameters a little bit better and even ask the tool sometimes, um put yourself in the role of of an instructional designer, what would you do, you know, in this situation? And and I think it it helps when you kind of give it that context, right? Of that as well.

SPEAKER_01 34:28

Yeah, I mean, there's some like basic prompting uh templates out there, you know, uh like you know, role, context, task, output is is like a really basic one, right? So if you you tell it what role it is, basically what you're doing is you're narrowing the statistical probability, right? So um you tell it you tell it the role, you tell it the context, right? Um, and then you you tell it what you want it to do, and then you tell it what output you want it to be in. Right. And you you'll you'll generally get if you can meet all of those, you generally get a bet better result um than just like a simple question prompt.

Jackie Pelegrin 35:01

Right. That's so true. Yeah. It's it's amazing when you start playing with these tools and experimenting with it. Like you said, until you start getting in there and using it, you really won't know what works and what doesn't for your context and your situation until you experiment with it and just give it a try, right? It's the best thing to do. Yeah.

SPEAKER_01 35:23

And I know some people that are not, you know, maybe are not, they don't think they're tech savvy. Um this is one of the advantages of these tools is you won't really need to be tech savvy. Right? You just you're just typing in text. That's it, right? You're not programming anything. You're not, you know, you don't need to know code. Um it's like typing a text chat, you know, and that's and then you can see, but then you start again, you start learning how to do it and get more efficient and better results quicker, right? So yes.

Jackie Pelegrin 35:52

That's true. Yeah. Yeah. We don't get better at something unless we we actually get in there, use it, practice it. Yeah, exactly. Yeah. I love that. So, Duane, as we wrap up, what advice would you give to instructional designers who want to embrace the AI without feeling

Adopt AI Without Rebuilding Workflow

Jackie Pelegrin 36:09

like they have to completely reinvent their workflow and what they're doing day-to-day?

SPEAKER_01 36:14

Um so I would say, you know, uh, some of the things we've talked about in this, you know, in this in this talk we've been having, um, where in those different phases of your instructional design, see where it maybe could help provide some different options to think through and you know, and and prompt for that. And again, that's just starting. That could be a pretty simple prompt, right? You don't need a lot. You can just and just see what it produces and you know, and then evaluate it, see what it does. And you know, maybe you'll get something you use, maybe something you won't, right? Um, and then just try try again with those. Um I I would I would caution trying to jump right into like some qualitative analysis and and and those and those type of tasks, because they can be um fairly complex. And we always want to make sure we're we're you know, we're we're keeping data. So like when we do qualitative analysis, you don't want to have people's students' names and you know, any of that personal identified information. So when you're doing that type of data, you make sure you want to clean your data before you use it in those systems. So that's why there's a few more pitfalls there, but for the brainstorming pieces of your design, I would say jump in and and and you know, ask it, you know, give it your what you're trying to achieve and um what you're thinking and see what it produces for you, and then you know, and then go from there, evaluate what it produces and maybe maybe you get some great ideas.

Jackie Pelegrin 37:29

Right. I love that. That's great. Just keep experimenting, right? And seeing what you can do.

SPEAKER_01 37:36

Yeah, and and understand, again, understand the limitations of the tool, understand the policies that are in place in your organization, right? You don't want to be violating any any you know um policies there, but you also want to make sure you're protecting information. Um you know, no person identifiable information and you want to make sure all that's because uh those open tools, you don't know where that data's going. Uh you know, absolutely. You want to be very aware of that.

Jackie Pelegrin 38:01

Right. And there's even like the settings too. I, you know, knowing you know what your settings, you know, are and I I've learned along the way. I'm like, oh, open AI, yeah, I gotta I gotta learn what their settings are. Exactly. You know, yeah.

SPEAKER_01 38:14

Yeah. Uh, you know, and that's you know, you start learning about the different tools, right? And the different things you can do. And like they have these things called temperature, right? You know, if it's hot or cold, right, it gives you different um uh rely maybe more reliable results. Um the different models, right? If if if you're doing something and you want it to be a little more uh sustainable over time, make sure you're using the same model, right? Right if you're running a different uh same prompt because you might get a little bit different result. Uh we talked about how previous chats influence your current chat. Well, you if you don't want that to happen, you might want to put a guardrail in there or a constraint. Don't use previous chats. Um I mean just just some things you have you start to learn um as you're as you're using the tool a little more.

Jackie Pelegrin 38:57

That's true. Yes. I love that. One of the things I like about um Chat GPT is that, and Claude has this too, where you can create projects. And I really like that because then I can go back to something when I'm like for this podcast or something like that, and I'm working on a project for that. It's nice to be able to go back to a previous chat and be like, okay, now I'm ready to work on this. Can you help me, you know, with with that? So it's pretty neat because then it has that history already there. So yeah, it's nice.

SPEAKER_01 39:26

Yeah, and I would say um don't take the name of the chat. So when you run a chat, right, it names it. You can rename it. Um I would tell you to rename it because you'll never find it again.

unknown 39:35

Right. Right.

Jackie Pelegrin 39:38

That's true. Yep. Good idea. Rename it. Yeah, I've learned that the hard way because then it's true, I can't find it. And I'm like, where did that chat go? Yeah. And then you're spinning your wheels trying to find it. Yeah, absolutely. Love that. Well, Dwine, thank you again for joining me and for helping us think about AI as a tool for innovation rather than

Where To Connect & Final Thanks

Jackie Pelegrin 39:59

just a replacement for thoughtful design. This conversation gives instructional designers like myself a practical way to explore AI while still leading with purpose, creativity, and care for our learners. In part five, our final conversation in this mini-series, I can't believe we're coming up on part five. Wow. We'll explore the evolution of AI and its ethical implications and content creation. So once again, Wayne, for anyone who wants to keep learning from you, uh, what's the best place to follow you or connect?

SPEAKER_01 40:28

Uh so please connect with me on LinkedIn. Uh, just do a search. Um, more than willing to continue a conversation there or or email me at my um National University email address.

Jackie Pelegrin 40:39

Great. Absolutely. Love it. And who knows, maybe some of your students will get to listen to this and and they'll go, I recognize who that is. Yeah, you never know. So yeah. Wonderful. Well, thank you so much, Dwayne, and look forward to part five coming up. Thank you for taking some time to listen to this podcast episode today. Your support means the world to me. If you'd like

Support The Podcast

Jackie Pelegrin 41:09

to help keep the podcast going, you can share it with a friend or colleague, leave a heartfelt review, or offer a monetary contribution. Every act of support, big or small, makes a difference, and I'm truly thankful for you.