In this episode of The L&D Explorers Podcast, we had the pleasure of speaking to Alexander Salas, an elearning expert.
Founder of Corpaxis, Alexander Salas, speaks on the rise of Artificial Intelligence in L&D spaces, and how industry professionals can make the best use of AI to create an effective training program.
Mentioned in the episode
Alexander Salas’ YouTube
Alexander Salas’ LinkedIn
Key Takeaways
1. AI’s impact on the L&D job market
While AI could never completely replace human insight in the elearning space, as it becomes more capable, there is definitely a shift taking place. Those who embrace and leverage AI in their L&D toolkit will fare better in the job market than those who don't. “It's going to make our workforce more competitive...the people that have experience and know what they're doing.” (6:52 - 7:14).
Actionable Steps: Instead of attempting to compete with AI, take advantage of it by learning to use it as a tool to produce higher-quality work. Gaining an intimate knowledge of AI in the L&D space and its shortcomings will help you remain relevant in the industry.
2. Increased efficiency
There is no question as to whether AI is more efficient at producing outputs. “[It can] spit out a video with an animated avatar and voiceover that is based on neural voices, all this happens in ten minutes. Now tell me what video editor or what instructional designer can do that, because I can't.” (17:36 - 18:05).
Actionable Insight: Use AI to expedite instructional design processes that are typically labor-intensive and prolong a project timeline. Not only will your team’s turnaround time improve, but they will operate in a way that is more cost-effective and efficient.
3. Implementing AI tools
As helpful as AI is, not every shiny tool will make sense for your organization. To assess what tools to adapt and how, Salas suggests, “...having an R&D sort of sub-team....looking at new tools, trying them out, testing them out, and… and then map… where can it serve, where can it be helpful." (17:36 - 18:05).
Actionable Insight: Before adopting a new AI tool, first establish your team’s needs and goals. Then, you’ll clearly be able to identify what gaps need to be filled and how AI can help you fill them. Otherwise, your team might end up spending too much time on a learning curve that does not benefit the organization in the long run.
We hope you enjoyed this episode of our L&D Explorers Podcast! Subscribe to our YouTube, Podbean or Spotify so you don’t miss the next episode.
Transcript:
Dan Gorgone: Hey, everyone, welcome back to the L&D Explorers Podcast from GoSkills. On today's episode, we're speaking with Alexander Salas, a talented learning experience designer and eLearning expert. And our topic today is AI automation and the benefits it can bring to your L&D teams. We'll discuss a variety of ways you can use AI tools to create courses and make your teams more efficient, as well as implementing these tools, questions about ethics with AI, and the trends we can look forward to.
I'm Dan Gorgone, course producer for GoSkills, and I hope you enjoy this episode.
We're here with Alexander Salas. Thanks for joining us, man. Alex is a learning experience designer and eLearning expert. He's worked with Fortune 100 companies and is very active at industry events, always supporting the instructional design industry. It's nice to meet you, man. Thank you.
Alexander Salas: Appreciate it then. Yeah, it's great to be here and thanks to GoSkills for having me here. Excited. Yeah, man.
Dan Gorgone: As far as today's topic goes, it definitely might freak out some of the people out there because it's AI. We're talking about AI-powered learning automation. And hey, look, honestly, AI is so powerful. I think plenty of us have heard about all the amazing things that AI has been doing and then we've also seen some of the kind of frighteningly impressive things that AI has been doing out there, like videos just sort of invented out of nothing.
But instructional design is absolutely impacted by AI and has been. And honestly, when I was looking up information about you as well, I saw, oh, there's an article that Alex wrote six years ago about AI impacting learning.So you know what's up, right? So let's talk about what is going on right now. How is AI impacting the instructional design industry and what kind of, what's your experience with it then?
Alexander Salas: Yeah, so, you know, it's. Thank you for mentioning that because it is true that the article that came around actually because I worked on the project as an instructional designer. So, you know, working as an instructional designer, as everybody knows you, it's a great job because you get exposed to many different disciplines and jobs and whatever the case. And at the time, I was working with Dell and worked with some experts who had a big notion of artificial intelligence, specifically machine learning, which is the true engine of this AI that we hear about.
But I think the one thing that listeners want to think about at this and something I was thinking about before as I was coming onto this episode and recorded the episode with you, is that notion that I was reflecting on and going, you know, when I was a kid, I'm a Gen Xer, so when I was a kid, we had all this talk about, you know, flying cars, right, spaceships or whatever, you know, like flying somewhere, like to the moon and automated refrigerators and what, all that stuff, right? But there was never anything that described what we're having right now.
So, like, there was never anything that said, hey, you know what? You can just put a sentence here, and here comes a video, right? Or, hey, you can just tell this one chatbot. I mean, chatbots by themselves, that wasn't a thing. I mean, we knew the Jets, right, as cartoons and stuff, but aging myself quite a bit here. But, you know, that's quite revealing, right? 'Cause, I mean, if you think about innovation, you can't have innovation without prior knowledge or something, so you can't think of. And I bet you to try, right? Because people say, no, no, that's not, there's, there could be something completely new that you never know.
This hasn't been anything completely new. Something is always founded from something else, right? But in this case, now we, we talk about this generative AI. I mean, that's what, what is it, what is it based on? I mean, it's based on the power, creating something, you know? So that's a challenge. So it's an interesting thing, and I think it's something that we should all be paying attention to.
Dan Gorgone: I definitely grew up reading, you know, Isaac Asimov. I grew up watching the Transformers. I- A lot of these things where, especially Star Trek as well, where you're thinking about. Think about what some of those writers and especially the designers behind those pieces of entertainment had to invent those things and try to invent the future while still being in the present and trying to imagine what would there be, what would we need and what would really make sense. And thinking about some of the tools that we have now, it's absolutely grown out of imagination, but also out of need, right? And when we talk about AI and the benefits that it can provide, absolutely.
Time is one of the things that has been a huge benefit for across all industries, but especially in instructional design. When you talk about putting in a prompt and saying, create me a course on this that is maybe this long, that reaches this level of difficulty for this audience and covers this, this and this, and then it does it. How far along are we in this process of almost eliminating people like ourselves from the process? Will we ever be able to eliminate ourselves? Because I feel like, no. Although we may trip, we may find ourselves transitioning into other types of roles where human beings are more valuable to the process. I don't know. What are your thoughts on it?
Alexander Salas: No, no, that's exactly it. That's a great thing. So to get better, the companies that are developing AI are going to need humans to train those systems, right? So, you know, you mentioned the article that I wrote, there are different models out there. There are supervised models, which is the most basic one, and that's the one that is training the system. Right? So like when you go to ChatGPT, for example, and you open it up and you say, hey, give me x and y and whatever, and it gives you an answer, and that answer is good, you can train it, you can click on the thumbs up and they will know, oh, okay, I gave somebody the right thing, which, you know, can work in different ways, but I think that is it going to eliminate people?
I think it's going to reduce, is going to, is a threat no matter what, it's going to reduce our, it's going to make our workforce more competitive. The job market is going to be way more competitive. So only, you know, we can, I think the folks that have experience is what I say, the people that have experience and know what they're doing, we can definitely take quite a bit of advantage of the situation and get ahead of the curve and provide that. Now it's going to take a little bit of suffering because this may sacrifice some of the things that we like to do the most. Right.
Like, you know, I don't mind sitting there three, 4 hours doing some edits and some videos, some composites and compositions, visual effects and stuff. I don't mind that. I find that fun, but in the business context, that can have a big impact. So I know if you saw recently, everybody says ChatGPT, and obviously ChatGPT is probably OpenAI is at the leading position of all of this in many respects, especially now that Google kind of had a flop with Gemini, Gemini or whatever, not being that good. But the Sora, which is something that's going to be available, that has super ultra creative, realistic video that wow is coming out of a few prompts.
So, yeah, right now, people using ChatGPT four can, and I've been doing this on my YouTube channel and CorePraxis YouTube channel where you can see that, you know, you can, you can leverage different applications and create something with minimal effort.
Dan Gorgone: Yeah, and let's dig into this a little bit too, because it's not just like, hey, create me a course that is like, the sole use of AI. And I mean, you mentioned Sora for generating video, but the variety of ways that AI tools can provide solutions for L&D teams, for instructional designers, it keeps expanding by every day. There's something new we're seeing some video creation, certainly image creation we talked about before we got on this interview. We talked about transcripts and translations and even a little deep fake of your mouth when the speaker is speaking a different language.
What else are we talking here? That can be an advantage and save time and money for the L&D teams.
Alexander Salas: Oh, yeah. Let's put in all the different perspectives. So the most accessible one right now for anybody is ChatGPT, right? So let's take that, because then there's that.
But then there's these other applications that people have developed using those models, right? So let's get specific now with ChatGPT and say the most accessible one, the one that seems to be free or very cheap, 3.5. And then there's four, version four, which is the latest and has the better tools. So if you don't know what you're doing or if you have no clue what you're doing, you say, give me a course on this. That is a duration and, you know, on whatever topic, right? What it is doing is looking at the Internet. So it's no different than having a person go look at the Internet and do that. And it helps if you have a master's degree, perhaps in a sense, that you know how to do research, right? That's the only skill that helps, because then you can sort the no good from the good, essentially. But otherwise, if you don't know what you're doing, you get that output, and then you'll be like, "Oh, great."
So now your learning objectives have to start with "understand," they're based on Bloom's taxonomy versus Mager's ABCD, because there are so many nuances in there. And when you write tests, when it writes tasks, if you tell ChatGPT to write you a task or a multiple-choice question, it's hard to write a good one, right? Observant practices, so that time can be cut down from hours to minutes. But if you don't know what you're doing and you tell it, "Hey, just give me ten multiple-choice questions, whatever." It would actually be true and false questions, which are not good because it's only 50% chance; it will give you all the stuff. So it gives you whatever it finds on the Internet. Now, if you are able to customize it, if you're able to give it instructions, have a conversation with it and say, "Hey, look, we want to use this framework. We want to use this type of model. We want to use this." Now, build me a course based on those parameters. Then you start getting some good stuff and then you can refine it.
But again, if you take that interface out and you just tell an instructional designer, "Hey, give me a course." The average person is going to take days to give you that. So with this type of tool, you can do it in ten to fifteen minutes.
Dan Gorgone: I started using AI tools in my course production, research, writing process. And, you know, to me, using AI felt like, alright, well, it's a tool. It's a tool. It's nothing to be afraid of. I know what goes into a good course. I know how to structure a course and put things together and understand the importance of learning objectives, all those things that go into a proper course. In my dealings with AI and giving it prompts, I started to realize that I was getting the best output, the best work, by treating it more like a collaborator rather than just a tool. Right. I would start giving it prompts where it would be very simple. "Give me an outline based on this." Alright, now we have this. Here's a learning objective for a lesson. "Give me a five-minute script that would be narrated," and so on and so forth.
But then, like, building upon that, and as you said, looking at it, refining it, going back and fixing things or adding things, using the AI tool like that, I found was incredibly valuable because it was almost like having another well-experienced person who just happened to have all the knowledge in the world in their head, but being able to give them those refinements that really make a huge difference and taking it from kind of like a base-level course to something that really is going to push the learner, uh, to really understand the concepts that we're trying to get across. Right? Right.
Alexander Salas: So we talked about ChatGPT in that perspective, right. Then you go to ChatGPT 4, and it has now this marketplace of apps integrated, right? So Canva, the graphic design tool; you have video. So Veed.io, another app. So those integrations are there. And what that does is now that you gave this input, you can give this input saying, "Hey, create me a two-minute video, or write me the script for a two-minute video that discusses emotional intelligence from a leadership perspective for the C-suite." Right.
And you might give it some more detail and they'll give you the script. Okay, now that I give you the script, you can follow up with that and said, "Okay, great. Now modify this so it can be set up as a prompt for video or whatever other tool out there." And now you do that. And now if you can spit out in the integration and spit out a video with an animated avatar and voiceover that is based on neural voices, all this happens in ten minutes. Now tell me what video editor or what instructional designer can do that, because I can't. And I've been doing this.
Yeah, so that's what I'm saying. Now, you know, we can put all kinds of things on and you can look at, well, but those avatars are not really that good. Or, you know, but they're getting way better and some of them look really good. So, I mean, when you're looking at environments like sales enablement, for example, where you don't have time to have a, I mean, a good tight development cycle for e-learning multimedia, level two to three takes about eight to twelve weeks. If you don't have that time, right, that's two to three months. Sales enablement cannot wait for that. So that could be an option to expedite that workflow.
Dan Gorgone: When you start throwing those numbers around, I think that L&D managers and directors start thinking, okay, this sounds like a significant money savings. This sounds like a way that we can restructure an L&D team and potentially get more out of less. But there are risks. You know, there's risks along with also, I mean, misconceptions about what AI can do, what you should maybe allow it to do. Like there's, you know, there's a balance that we have to reach. And as AI evolves, I mean, it's going to continue to change.
So people need to keep on kind of tweaking things here and there to make sure that they get what they want. So if you're an L&D manager out there who's hearing all this and probably has read a lot of things and heard a lot of things already, what's some good advice for an L&D team as they're thinking about implementing AI tools? Is there a way that they should start? Is there a first step that they should be taking?
I would have to think it's something that is internally looking where they're looking at their team and looking at their needs, or maybe even looking at their budgets and trying to figure something out but what do you recommend here?
Alexander Salas: Oh, yeah, that's a great question.
You know, from a leadership perspective, what I recommend, you know, one of the things I always pushed in many of the teams that I work with, I always push for having two things. So one is set a lot of time like, you know, an hour or two a week that is based on professional development, right? So professional development is a lot that could be like a Thursday morning, whatever, Friday morning, whatever the case. And then the other one is having an R&D sort of sub-team, so, meaning research and development, not learning and development, but like a little bit of a research. So folks in your team that like doing the stuff, right, looking at new tools, trying them out, testing them out, and then having those folks actively on a, on a weekly basis, monthly basis, try out different approaches and different tools and then map to what is that can be needed, where can it serve, where can it be helpful?
So, for example, if I was going to do that right now and you and I were in a team, I already know many of the different tools, right? And if you were like, you know, the two of us are peers and the managers, like, hey, guys, can you find, how can we make our process better for video and this and that?
Now I'm going to tell you, well, you know, I know that Dan and I can go to Adobe Premiere Pro, pump in a video, hit transcriptions and, you know, transcription and generate a transcript and then generate captions that, to burn captions that look kind of cool on the video. That takes 2, 3, 4, 5 steps, right? You got apps out there right now that you can just upload a video and it will do it. And it will not only do it, it will animate the captions to accentuate with contrast the words that are being spoken.
Aside from that, it also has the capability of publishing the different formats and the different specific formats. If you're doing social media internally or if you're doing landscape portrait, whatever the case, you know that it's got to be different if you're going to Instagram or if you're going to YouTube or if you're going to TikTok. So there's one solution there already that is going to at least save a quarter of the time that it takes to make video. So, you know, that's something the leaders should be looking at is having someone in their team kind of looking at these things ahead of time, seeing where the value is, get those notes down and then you won't be the traditional thing, which is like the business side of the thing thing. You know, the operational side comes to you with the idea, hey, have you guys seen this? Have you seen this? Right? And then one day is like, oh, there you go. Take a look at that. Because it's pretty good, right? Yeah. Right.
Dan Gorgone: The shiny new toy that, that comes out like every other day. Right?
Alexander Salas: And then you can look way better in the current organization. You'd be like, yeah, we've been doing it for a minute. Or, yeah, we looked at it. It's not really that good. It has its problems. Or, yes, we looked at it, it's great. When you want to start, you know, it's a different conversation. I think it changes everything, too, because it also changes your positioning in a way that you can get better leverage when you now want to bring toys, you want to bring in for your specific purpose.
Dan Gorgone: We talked about, like, how quickly things are evolving and things are changing, and with that comes, you know, more risks. And one of them that is definitely in a lot of people's minds is the ethics behind AI and being able to ensure an ethical use of the technology and especially the data that it's pulling from. And you talked about machine learning and how these AIs will take data and take information and take inspiration, shall we say, from so many different sources out there, whether it's text-based or whether it's artistic base. And I had a great debate with my son, who was an artist, about the nature of AI art and what it all means and what truly is art and what's not. And it's fascinating to hear what he and many other people out there who have really created stuff on their own have to say about AI.
So I put it to you, as far as instructional design goes, do we have worries about the ethical use of AI when it comes to education and structural design? What are they?
Alexander Salas: Well, I think primarily, yes. I mean, always the one thing is you should never use this as a plug-in and we're done type of thing like we discussed. You want to look at it, refine, you want to also have some kind of input, some kind of correction to it. Right. And once you add that, then you're transforming that piece of work.
So in one perspective, so if we think about it from the legal perspective, I'm not a lawyer, but I've done quite a bit of look into copyright because that's the main thing we want to be concerned about. And so here in the United States, there are different things you can do. So obviously, you comment on the YouTube video and you add enough value so the different categories in there, which makes, you know, I can use somebody else's video comment on it, and if I'm adding value with my input and everything else to it, it becomes something new. So now it's, it kind of falls on the fair use, give or take, right?
So in this situation, kind of same thing. ChatGPT is aggregating, which is basically what an instructional designer in most cases does, right. If you talk to a SME, three SMEs, you get input from three SMEs. Where's that input coming from? Right. If it's very specific because they do a specific job and task. Okay, okay. Maybe if it's completely proprietary. But when you get into the situations, of course, development for leadership and things like that, well, technically speaking, you should be giving credit to Dan Goleman for emotional intelligence stuff. You should be giving credit to Gardner for multiple intelligence work if you include it in yourself. So, I mean, I don't think it changes much. You should definitely do some editing, some value changing on it, some input to it, not just take it as it is.
But now let's consider that many of the popular apps, because when we talk about L&D, there's a circle, right?There's a village. I call it the village inside the village, you have the tools that L&D talks about, but that doesn't mean that those are the only tools. So in this case, when you go into the village, you have apps out there that have already added integration or integrated some way of ChatGPT into their workflow so they can say there are AI, because let's not forget that it's a selling point. Right.
Dan Gorgone: Well, I want to bring our discussion to a close by mentioning once again that you wrote about this years ago and you had the foresight, and so clearly, you know where things are going. You're a fortune teller, you're a wise man. You know where we're going. And that's where I want to go right here with the end, which is what emerging trends are you seeing? What do you, what do you think is going to perhaps impact our jobs as instructional designers and the L&D teams out there who are looking for solutions that are going to make their jobs, their work, more efficient, cheaper, better, help the learners, help us, you know?
So what are you seeing out there?
Alexander Salas: Well, I mean, you know, I think the forces that are going to change, whatever is going to change for us is going to be based on the same things that were forces before. So if a business, the people that pay your salaries, right? So if you work in a company and there are people that pay your salaries, know for a fact that what you are doing can be ten times more efficient and ten times, you know, the same quality and better and cheaper and you're not taking the steps and you are not showing any, you know, inkling of that, then, you know, that's when the layoffs come, right? That's just business. It's just what it is. We don't like it, but that's what it is.
So the other part of that then is to consider in what I leave for people. There is like we had this discussion primarily is, look, people say, well, AI is not going to replace you. Someone using AI will. And that may very well still sit as a legitimate, legitimate thought. If I have a, if I'm hiring people, there's a pool of ten people and three instructional designers are using AI and everybody's pretty much comparable in terms of the skills. It's a no-brainer.
We're going to talk to the people that are doing AI because there's something extra there. There's something just like the same was, you know, five years ago, talking about having a portfolio, an online portfolio, and there's still people going like, oh, no, everything. I don't. Proprietary and I don't have a website. I'm like, okay, well, you know, don't complain when you don't get called and you get selected. So the one thing I leave everybody thinking about is, look, the job market is going to shrink.
Does it mean that IDs are going to get done is over? No, what it means is you're going to have to work somewhere else is, you know, the corporate setting is going to be reduced to people that are working with these tools or are elite and what they do and then the other settings that have been behind for years, education being one government, right. All the settings are still going to be open for that. But then that means that that pool of people will start going over there to kind of get a job on that.
So, I mean, if you want to be safe, I will say, if you don't want to deal with any of this technology-wise and you want to be safe, go get a government job, right, until the government catches up. And then you have to catch up. But besides that, I mean, I will say there are tools today that will create a full course like you see in the popular tools with a few prompts. And all you have to do is edit a couple of things. We talked about video, we talked about audio. So what else is there, right?
I mean, it's something to consider.
Dan Gorgone: So educating yourself about the capabilities of AI, understanding where things are going, getting plugged into the community like you talked about, it's only going to benefit you as an ID as well as a, you know, a manager of L and D staff of someone who is providing that direction and certainly trying to organize and manage the tools and the people who are producing those things.
So not only will, not only can AI create better things more efficiently, it's going to help you keep your job as well. So, Alexander Salas, I want to thank you very much. This has been a great talk and learning more about AI.
I'm trying to educate myself just like everyone else watching. Got to do it. Where can people learn more about you, man?
Alexander Salas: Well, particularly, you know, right now, working on our project, which is the YouTube video, the YouTube channel called Corpraxis.
So c o r p r a x is corpraxis.com.
It will be a platform that will be available for folks to learn about this stuff and then the YouTube channel. I'm sharing a lot of the tools already that I've been talking about in this episode. And aside from that, then, as you guys saw me on LinkedIn, there are quite a few followers out there, and I'm pretty much posting every day there on different topics. So yeah, reach out. Just let me know that you heard me here in this episode and we can connect.
Dan Gorgone: Hey everyone, thanks for watching this episode of the L&D Explorers podcast. If you enjoyed it, please give it a like and subscribe because more episodes are on the way and no matter what your learning and development goals are, GoSkills can help. Click the link in the description to find out more.
And thanks again for watching.
