Learning and development professionals are finding new ways to leverage the power of artificial intelligence every day. Michelle Echevarria, director of learning and development at JBS International and an early champion of AI, shares some best practices for using AI in the L&D industry.
Michelle Echevarria offers many ideas for using AI in your daily L&D work including the following:
Powered by Learning earned an Award of Distinction in the Podcast/Audio category from The Communicator Awards and a Silver Davey Award for Educational Podcast. The podcast is also named to Feedspot's Top 40 L&D podcasts and Training Industry’s Ultimate L&D Podcast Guide.
Learn more about d'Vinci at www.dvinci.com.
Susan Cort: [00:00:00] Joining me today is Beth Buchanan, Senior instructional Designer/Project Manager at d'Vinci and our guest, Michelle Echevarria, Director of Learning and Development at JBS International, who is joining us to talk about her AI journey. Thanks for coming on the podcast, Michelle.
Beth Buchanan: Welcome, Michelle
Michelle Echevarria: Thank you so for having me here.
Susan Cort: Michelle, start off by telling us a little bit about J B S International and your role there.
Michelle Echevarria: Absolutely. So I've been at JBS International for about 10 years now. And currently I oversee the Instructional Media Solutions Group. Essentially we provide, um, services and solutions, around instructional design and development for our, external clients. Uh, the vast majority of our external clients are in the public sector from a myriad of different, uh, government agencies. , JBS has been around for quite some time, over 30 years now.
Michelle Echevarria: We have a deep, healthcare, social sciences, education, and workforce development [00:01:00] expertise. Um, we work to advance the, fields like health. Safety, um, you know, keeping in mind and wellbeing of people, uh, focused on where they live, learn, work, and play. And, you know, we, pride ourselves on, uh, our, our deep level of expertise and in-house SMEs that, you know, definitely works to set us apart, from other agencies such as us.
Susan Cort: Well, that's great. Well, we met you at the Training Industry Conference and Expo and had such a great conversation about artificial intelligence. I knew in about 30 seconds that we had to get on as a guest on our podcast because you, are blazing the trail. You, you are, you are a few months ahead of most of us and I was just so excited to talk with you.
Susan Cort: I thought it'd be great, uh, for Beth and I to have you on the podcast to learn a little bit more about your journey.
Michelle Echevarria: Absolutely. Well, like I said, I'm excited to be here, uh, and excited to share, uh, my knowledge and, and what we know so far.
Beth Buchanan: Great. Well, let's kick, let's get [00:02:00] right into it then. Um, I think we both know that AI is absolutely everywhere. Um, it jumped on the scene eight or nine months ago, it seems everybody is doing it now. So big question is how can AI be used to improve learning and development? From your perspective, how can people use AI to their advantage in this field?
Michelle Echevarria: Right. Um, I think in respects it's still very much, uh, an open, uh, answer for that. Uh, 'cause right now we're still very early in these, AI developmental stages. It's, it's sort of the, uh, wild west of the AI phase right now, but tools like Chat GPT, which is, I'm sure, all of the, or at least I hope all of the listeners at least have, uh, heard of, uh, is an example of something that could be used within your, um, L&D processes relatively easy and quickly. It's a free tool and for example, it can create [00:03:00] or help you create things like your SME interview questions for your analysis part of your ADDIE model, presuming that you're using that. Um, it can also help you develop topics for your training.
Michelle Echevarria: It can also, work towards, uh, creating an outline. Um, and in some regards even create sort of earlier drafts of, uh, your complete content that you want to include in the module. So it's, it's pretty robust. Uh, it can be used in a lot of different ways. Um, like for example, I anticipate tools like Articulate and Adobe, uh, e-learning products. Soon they're going to start encompassing all of these AI tools, such as a ChatGPT, and they're going to roll that into their software, um, as part of, just as, so it'll be sort of a normal part of our process, uh, in developing pretty soon. I'm sure that they're working on that [00:04:00] right now.
Beth Buchanan: Absolutley. Yeah. So people should not run and hide.
Michelle Echevarria: Absolutely not.
Beth Buchanan: This new technology, right? I heard a few things that you said that I thought were really interesting. Um, know, using it kind of as that first step project, right? Whether it's researching or outlining or, know, just coming up with a topic you could say, could ask ChatGPT, Hey, give me 10 topics, you know, larger topic me started, right?
Beth Buchanan: Um, do you think that's helpful? Like do you think that's the first step in producing quality learning? What do, what do you get as, you know, someone who develops content? What do you get by using this tool to start out?
Michelle Echevarria: ... um, in terms of creating quality learning, uh, in its current stage, I think not so much right now, humans, um, still sort of reign supreme in that space. Uh, but it can certainly help to create sort of a level one draft of things like an outline, uh, things like [00:05:00] script. And storyboards things like that. But it would still need to have a human oversight, whether it be from, uh, an instructional designer and or a SME. But what it can do is cut down on some of your upfront, development time so that way you can focus on creating more robust content.
Beth Buchanan: I love that. I think everybody loves that if they're in this field. Right? Any kind of space to create more, to gain more efficiencies and time ,right? So what are some ways that you are using it personally your day-to-day?
Michelle Echevarria: Personally I started to, uh, sort of play around with the, um, AI around the, uh, holidays around Thanksgiving, Christmas time. So I actually used it to help me write, uh, a couple of holiday notes to people, um, on a personal level. It can also help with, uh, crafting emails. and then I've also. [00:06:00] Uh, heard of people who have used it for, uh, even medical items where you may need to write an appeal letter, uh, to your health insurance agency. So it can help with some of those sort of everyday things that we need. Uh, even things like recipes, workout regimens. So there's a lot of really creative ways, uh, you can use it in your personal life as well.
Beth Buchanan: Yeah, that's exciting. I haven't even thought about that yet. I was, know, that saves me a little time writing Christmas cards, that might be…
Michelle Echevarria: Well ,of course you want to go back and personalize it a little bit more.
Susan Cort: No.. Now, Beth, what you and Michelle could do is use ChatGPT to ask it how it could help you, and it would tell you so
Michelle Echevarria: That's true.
Beth Buchanan: Exactly, Um, apart from that, how is your organization using it? How are you using it at work and how is, you know, you know, sort of a personal day-to-day for you, but [00:07:00] then larger picture. How's your organization using AI as well? Sure.
Michelle Echevarria: Um, I think in, in, in sort of two uh overarching ways. One is in, on the production end, they have a lot of really good tools out, uh, that can really help fast track the development process, like, for example, Adobe Podcast is a really good tool. Uh, like example, a pod podcast that we're doing right now. Uh, podcasts has become an, uh, a tool that's increasingly been used in the L&D space, uh, whether it be scenario based to interview SMEs, et cetera. So, and a tool like Adobe Podcast can help you. Uh, cut down, uh, exponentially on the editing time. It can remove silence, it can, uh, remove background noises, um, and also remove the, um, just what I just did right now -ums and [00:08:00] ums.
Susan Cort: The non Yes, all, all the non fluencies. It takes it right out.
Michelle Echevarria: It takes it right out. And then it also just overall enhances the overall quality of the audio output that you had, get it closer to studio quality, uh, especially with the advent of, uh, recording more so online and virtually, pre pandemic. Uh, you could go into a studio, uh, you would record um, in a, with using professional equipment. Uh, but now, you know, post pandemic, , you know, for our clients for example, uh, they've really got accustomed to recording remotely. It's a lot more comfortable them and you can use a lot of different tools for that, like the online tools, Zoom, MS Teams. Um, but on our end, once we get the, the product, you know, putting it through a software like Adobe Podcast can really save a tremendous amount of time for us. And then also, ChatGPT can really help [00:09:00] with, um, idea development and brainstorming. Uh, you can create a prompt like we talked about a little bit before, um, where you can, uh, go back and forth with the ChatGPT, uh, uh, product on any specific topic that you have. You can even go so far as to, give it a bit of a personality. Like for example, if there is, um, let's talk about philosophy you can ask it to pretend to be Socrates , uh, yes, absolutely.
Beth Buchanan: Really?
Michelle Echevarria: Pretend to be Socrates.
Beth Buchanan: Oh, I didn't.
Michelle Echevarria: Right? And then you can, you know, you can propose, you can make a statement and you can ask it to pretend to be Socrates and essentially have a conversation back and forth with ChatGPT.
Susan Cort: I was just thinking that isn't that one of the ways you could really, uh, practice kind of, um, some soft skills, you know, not, not by pretending you're talking to [00:10:00] But by giving you the opportunity to, to practice that way with, uh, with a tool like a ChatGPT
Michelle Echevarria: Absolutely. Um, interviewing skills, uh, it can create interview questions, uh, for you, it can give you, what would be ideal answers to questions. You can ask it to elaborate. can ask it to, uh, uh. Uh, poke holes in your, your theory or your idea. Um, you know, pretend to be a senior executive at X,Y,Z company. Here's my proposal. Now, ask me these questions as if you were a senior executive. That um can help you prepare for presentations and, uh, just really, you know, help you really solidify what it is that you're trying to do.
Beth Buchanan: That's really neat. I haven't, I haven't tried anything like that. I just, I, I find it fascinating how conversational you can be with ChatGPT, right? You give it [00:11:00] a prompt, gives you something back, and then you just, just kinda say, no. What about adding something like this in? Or what about changing it this And it's just this iterative process, just like a revision process, but, It's with
Michelle Echevarria: Right.
Beth Buchanan: bot, right? That's just you this, you know, feeding back to you, generating for you
Michelle Echevarria: Right. So, you know, ChatGPT, um, the g p t stands for generative pre-train transformers. So I think that's where the, the, the pivot point came because, you know, AI machine learning is, is not new, um, at all. But what is new is for your regular person, your, your machine learning layman to be able to communicate. And in some respects be able to control, um, the interaction, what happens there. So when that happened, that's when this really exploded because then now everyone can use it as opposed to, previously you had to be like very technical or you worked at, in an academic space [00:12:00] or in a, uh, military space or something like. Now it brought it to, to the masses.
Beth Buchanan: Going back to, um, your company, J BS International and sort of how is your organization using it, any kind of AI, broadly speaking, uh, to help improve the L&D initiatives there.
Michelle Echevarria: In terms of the, the learning and development, uh, space, uh, primarily we've found a lot of, usefulness around the, uh, sort of production end of it. And for the, in strictly instructional design piece, uh, the analysis piece, uh, helping us to, , flesh out topics, uh, Questions. To be asked to the SMEs. Um, helping us sort of like broaden our, uh, scope of what the possibilities could be around, uh, the training [00:13:00] and different training modalities, um, has really been useful to us.
Susan Cort: How have you seen that work? I mean, how successful has it been in getting your team members on board and, and really making a difference in the work that you're doing?
Michelle Echevarria: So in terms of getting the, the team members on board, luckily, uh, uh, working with my team in particular, we've been, we were sort of the, the early adapters of the bunch. Uh, so . As early as early this year in January, we started to loop it into what it is that we do, um, from, again, the, uh, content creation piece and then the production piece. Um, on the production end. Um, in, in some certain tasks have been cut down. Let's say for example, uh, took up to . You know, four hours to do, about half a day to do. Um, uh, we've been working with tools can help with that. Us cut that down in some, [00:14:00] uh, cases, even to a few seconds. Um, so it's literally a click of a button. Um, so as you can imagine, uh, it's been an incredible boon for us, um, around that. And we're always learning, you know, always open to, um, you new ideas and, and new uses of AI. Um, and just, you know, trying to stay in that, uh, almost having a sort of like an innovative kitchen kind of a feel to it.
Beth Buchanan: Hmm.
Susan Cort: That's great. I love that because it, this is not a set it and forget it thing. It's not like you're doing AI and you will never continue to iterate it. I mean, you're going to be experimenting every day, every month looking for new tools, new uses. I mean, you're never really done. Right?
Michelle Echevarria: No… I mean, there, there's, there's so many tools coming out every day, and it was come, they were coming out at light speed earlier this year, and now that, you know, uh, I, the mainstream is caught on, it's even more so. [00:15:00]
Susan Cort: Yeah, now they're, now they're going to start buying each other up,
Michelle Echevarria: What's gonna happen, right? you I talked about, you know, early, like, uh, Adobe, uh, Photoshop now has embedded, uh, AI, prompting. Uh, they're, you're able to remove, remove backgrounds. Uh, automatically when before you had to use an external tool for all that, an AI tool to do that. So they've already started either implementing, uh, and as you mentioned, they'll likely start buying up all of uh, different AI
Susan Cort: Mm-hmm.
Beth Buchanan: One thing, one thing I would, comes to mind though with, with the rapid pace that everything is exploding right now. you know,, we have to keep in mind some sort of safety around the whole thing, right? I mean, there needs to be some general guardrails that we establish maybe personally or within our organizations, um, they don't really exist yet. What, what have you guys, know, what, what do you have to say about that piece
Michelle Echevarria: In terms of [00:16:00] guardrail, one universal rule should be to never, never, never include your personally identifiable information things like your address, where you work, any medical information inside any of these tools. . Keep it as high level and generic as you, you possibly can.
Michelle Echevarria: For example, if you are, uh, creating a, a prompt that's specific to your company, uh, your name of your company is Momentum Interactive, you can do something as simple as, uh, do a search and replace and, uh, Momentum Interactive, let's say for example to, you know, ABC company. Or something like just to clean it make it as generic as possible because again, it's uh, it's ChatGPT, so it's generative pre-trained transformers, so it's always learning.
Michelle Echevarria: So you [00:17:00] want to make sure that it doesn't learn from your proprietary business, because in theory, someone, uh, in another space could be working up a, prompt and it could inadvertently glean from the information that you've now input into the program. So that's something that you wanna be, definitely careful about. And then always implementing, uh, a human checkpoint, is another thing. You, you never want to go into these tools and, you know, just take it for face value. Like for example, some of these, , tools like a Mid Journey and things like that, uh, anyone who's been working with with graphics know that AI has problems with hands, don't do hands well, which is a weird thing. Yeah, it's really . Um ,um, so I, I've, I've seen, uh, folks use AI images, um, [00:18:00] and the person has, two pairs of hands. So they have four hands sort of odd things or like more fingers than they should. It's really, it's really interesting and it's something that if you're not looking careful, like you may miss, um, and you know, if you've been, been working with the tools and it's not just Midjourney it's about of them has, has have an issue, uh, with hands.
Michelle Echevarria: So that's something That's small. Um, but it can, you know, it can, uh, create an issue if you were to deliver this to a client and just, you know, being mindful of trusting the information. 'cause um, some of these programs can do what we call hallucinations, which essentially means can make up content and information, things that have never happened.
Susan Cort: Well, it's just, it's just a good reminder for people to, you know, use all these tools with enthusiasm, but with caution that it's not, it's not meant to replace everything we do, but there [00:19:00] are some definite good use cases. Like you were saying, Michelle, that uh, you know, things that save time, that, that give us the space and time in the day that we need to be more creative or more thoughtful or, or, you know, uh, just more efficient.
Susan Cort: So I guess just be careful, right?
Michelle Echevarria: You know, be very cautious. Um, I've, I've, even myself, I've asked it for, you know, uh, evidence-based results and to provide the links. And the links. Sometimes they don't lead anywhere. Or they're not accurate. So you, like, you always need someone to, to check it out. And, you know, in addition to the hallucinations, you know, as a part of your guardrails and your safety, it's also be mindful of bias, biases, uh, in everything. Anything that a human touch has some level of bias. So that's also something to be mindful when you are prompting and then when you also get the results from your prompt kinda keep that in mind and has some checks and [00:20:00] balances, um, you know, most, like we mentioned before, most of the information which was gathered from the internet, whether or not it was actually vetted. Is another remains to be seen , you know? Um, uh, so, you know, for example, in, in, an example in the, um, uh, the EU, you know, one of the things that they're trying to do is, um, require these AI firms to list where the training and data that they have actually came from.
Beth Buchanan: Yeah that's something I've wondered about a lot because, you know, how do you cite your source? And, and do you have to, is is the assumption now with AI that know, everything is fair game, like open sourced and we don't need to, to actually, you know, come up with a list of citations every time we develop something? I mean, where do you see
Michelle Echevarria: Um, I think the lean is towards actually citing, you know, where you got it from and whether it be a Chat GPT kind of a thing, um, [00:21:00] as opposed to, because otherwise that, I mean, you didn't really, you didn't technically create the content. But my hope is that, you know, the, the information that you do get from these types of tools that you, you're vetting it and you're updating it.
Michelle Echevarria: Like what you get from the tool is level one. Right, so it's not level three. what you're releasing to the public, to your client, especially to your client, should be level three. It's been vetted, it's gone through an editor, et cetera, et cetera. So that, that really shouldn't be an issue, hopefully.
Beth Buchanan: Do you have conversation? Do you have of an open conversation about it with your clients? As I know a lot of listeners are in the same field where they're working with clients and maybe don't know how to approach saying, Hey, I wanna use AI to enhance my work, but gonna be comfortable that? Like how do you, how do you navigate that conversation?
Michelle Echevarria: Well, in our case, we've had a couple clients that have released, uh, statements and or sort of interim policies around AI. And, uh, what we've seen so far [00:22:00] is essentially, uh, you yes, you can use them. Um, we understand that it could potentially be a tool for, uh, innovation and, uh, we encourage, uh, exploration and experimentation, but it shouldn't be you know, the way that you are developing, you know, content on our behalf, like we talked about before, the vetting and the human checkpoint should always be there. Um, and I think that's sort of the general, of uh, perspective that, a lot of companies that should be taking this should be, , the time of exploration and experimentation. To figure out, you know, what do you need, what kind of tools that work for you, what process works for you, what kind of security measures work for you? Um, is definitely, this is the time for that to start working on that and developing that.
Susan Cort: Michelle, I know you and your team are, are still navigating all this of course, but you're perhaps many months ahead of, of [00:23:00] some others in the industry. For people who are just getting started using AI or trying to figure out how they might use it in the L&D industry, you know, what, what would you tell them about how they would get started and just some things to look out for to make sure they were, um, kind of navigating this journey properly?
Michelle Echevarria: So first off, for anyone, uh, 'cause there are people who are are who sort of fearful, right? And scared and, you you watch these movies where AI sort of takes over the world and, you know, all these kind of things happen, . Um, so the first thing is to know that everyone in some way is already using AI. You just are, uh, whether it be when you're texting the predictive text that comes up, that's AI you're watching Netflix, and it suggests movies that are relative movies that you've watched previously. That's AI. Um, in the L&D world for, uh, programs like [00:24:00] Storyline and Captivate, they have a text to speech function that's AI so you're already using it. So I would say that again, this is, we're very early on, uh, in these stages. Everybody's still learning learning. No one say unless you've, you've been in machine learning, you're a machine learning professional. Uh, no one has their 10,000 hours yet. Everybody's learning here at the same time. Um, so this is the time to experiment figure out what works for you, works for your team, uh, and then start to implement it. Um, in terms of in overall implementation, you know, you can have sort of level one, level two, uh, level one, uh, may be that your, your company just isn't using it yet. Uh, level two be this sort of experimental phase you're working through. Level [00:25:00] three. You figured it out. You have a, a pretty good grasp on the types of tools or that you want to use. Um, you have some general policies around using AI then label level four - it becomes part of the DNA your company where it's absolutely seamless and it's just something that is expected to be able to, to use on, in all verticals, in all departments, in all spaces. Um, a part of your development process.
Beth Buchanan: That was a lovely answer and I love, I love that you provide a framework, right? . Don't have to jump in with both feet if you don't want to just take a take, you know, dip your toe in and just, know, start with level one and let yourself get comfortable, use the tool and then, know, grow with this new idea, um, you.
Michelle Echevarria: So to help you get started, you get I would suggest go to Chat GPT Open AI, up a free account. [00:26:00] a hundred percent free, right? So in your first chat that you're gonna open, you can write a very simple prompt. So I'm gonna lay it out for you. I'm a senior instructional designer and I want to create course on. You can insert topic here, whatever the topic is. Uh, I'm gonna choose the best practices for safety in an in-home daycare. Give me five potential topics for the course. You press send, gonna come right up just a few seconds. Not even a second will come up and then after that, The second prompt could be something like, and this is using the topic that you choose, you are gonna take best one. Give me an outline for three modules on health and hygiene protocols, [00:27:00] it's gonna come right up. Second two seconds. It's gonna write it all out for you all in that same chat. Uh, I think you'll be blown away by how quickly it is and how it can give you a framework level one, depending on the generality of the topic. health and hygiene is pretty generic. Uh, it can even get you to level, two.
Susan Cort: It gets you, it gets you thinking, you know, if, if nothing else, it, it's not gonna do your work for you necessarily across the board, but it, but it'll get your, your juices flowing and it will get you thinking about maybe some things you hadn't thought of before.
Michelle Echevarria: Absolutely, absolutely. It's a partner. Um, it's not gonna take your job. Um, but it may if you don't learn how to use it, so you, you're gonna have to partner with tools. It's your assistant, it's not your boss.
Beth Buchanan: I'm struck by that last sort of, foreboding, uh, comment it may if to use it. So just, um, you segued perfectly into my next question, which [00:28:00] was what do you think is the future here? What do you think is the future of AI and learning?
Michelle Echevarria: I think eventually will become sort of a clandestine partner and powerful tool, I think we talked about a little bit earlier in, you know, our everyday softwares that we use, you know, Storyline, Captivate, or whatever else comes 5, 10, 20 years down the road. It'll be wrapped into the DNA of those tools almost to the point where we won't even know that we're using it. It, won't even be highlighted as an AI tool. Like everyone has AI tool now wrapped around there because it's, it's, yeah, it's sort of the hot right now, but it's gonna be included in the DNA of all the software tools. And then also in all the development processes. 'cause by that time, people will have included it. As a part of their general business practices. Uh, it won't be this sort of new shiny thing anymore. Something internet, new and shiny Right. In the, know, early [00:29:00] nineties, mid
Susan Cort: or, yeah, or Or social media, which was also new and shiny. Yeah.
Michelle Echevarria: And now just something that we just do every day. It's incorporated to our everyday life. Um, and, know, I think that's the future uh, AI tools in L&D.
Susan Cort: Well, there's so much more for all of us to learn. We'll have to use AI to continue to learn about AI. So any other final points you wanna share, Michelle?
Michelle Echevarria: I do wanna encourage people, to, you know, start to dip your proverbial toe the water of AI. If, if you haven't, the water's warm. This is the time to experiment and, and figure things out. Figure out what's gonna work for you. And it's an extremely powerful tool and once you learn to harness it, possibility is endless.
Susan Cort: Great advice. Thank you very much, Michelle.
Michelle Echevarria: Thank you so much for having me.[00:30:00]