Beyond Course Generation: What AI Still Doesn't Understand About Learning

00:00:10 RK.Prasad

Good morning, good evening, wherever you are.

00:00:12 RK.Prasad

Welcome to this very, very interesting podcast and on a very hot subject that is artificial intelligence.

00:00:14 RK.Prasad

The title of this podcast is What AI still doesn't understand about learning.

00:00:21 RK.Prasad

We all know AI can build courses, but can it design learning and develop better designers and better learning?

00:00:32 RK.Prasad

Over the last 2 years, AI has made content creation dramatically easy, easier.

00:00:39 RK.Prasad

Today, AI can generate objectives, storyboards, assessments, scenarios, scripts, videos, and even a complete course.

00:00:49 RK.Prasad

But this raises a deeper question, if content, creating content is becoming easier.

00:00:55 RK.Prasad

Why do so many learning programs still struggle to change behaviour and improve performance?

00:01:03 RK.Prasad

The other question is it making instructional designers job easier, better or is it in its own way deteriorating the output?

00:01:20 RK.Prasad

Perhaps contract was never the bottleneck.

00:01:22 RK.Prasad

Perhaps the real challenge lies in analysis, design, judgment, performance thinking, and understanding how people learn.

00:01:32 RK.Prasad

To explore that question, which is now very critical for the learning fraternity, I am delighted to welcome Ben Kanspedos, who brings a unique combination of.

00:01:47 RK.Prasad

Solid instructional design expertise and hands on work on AI enabled workflows, having worked in very large Fortune 500 companies in the United States.

00:02:01 RK.Prasad

Ben, welcome once again.

00:02:05 Ben Kanspedos

Thank you, RK.

00:02:06 Ben Kanspedos

It's great to be here.

00:02:08 RK.Prasad

Yeah, let us start off with question with the first question, Ben.

00:02:14 RK.Prasad

I'm just repeating what is there in the introduction.

00:02:18 RK.Prasad

AI has made content creation dramatically easier, yet many organizations still struggle to create learning that changes behavior.

00:02:27 RK.Prasad

What does that tell us about where the real bottleneck is?

00:02:32 Ben Kanspedos

Yeah, that's a great question.

00:02:34 Ben Kanspedos

And the way I see it, AI can do two things.

00:02:37 Ben Kanspedos

It can make training faster.

00:02:40 Ben Kanspedos

And it can make training better.

00:02:42 Ben Kanspedos

And if you don't nail the second one, then you're just creating training, bad training quickly.

00:02:49 Ben Kanspedos

So I've been in L&D long before generative AI was a big thing.

00:02:54 Ben Kanspedos

And from my experience, it's time was always the bottleneck.

00:03:00 Ben Kanspedos

And just putting the courses together, the content, putting everything on the screen, the images, I mean, that took most of the time.

00:03:08 Ben Kanspedos

So there really wasn't.

00:03:10 Ben Kanspedos

enough time for the instructional design aspect of it.

00:03:15 Ben Kanspedos

But now AI has pretty much taken care of the content development.

00:03:21 Ben Kanspedos

So it's really about how do you use AI to make the training better?

00:03:27 Ben Kanspedos

And like you were saying, how do you make it so it actually influences behaviour?

00:03:33 Ben Kanspedos

And it's really the instructional design that's the bottleneck.

00:03:36 Ben Kanspedos

And I'll give you an example.

00:03:40 Ben Kanspedos

You could have a course that maybe normally took 40 hours and now AI can reduce that to one hour and that's great, but that training might go to 1000 employees, maybe over the lifetime it might, you know, 10,000 people might take it.

00:03:58 Ben Kanspedos

And if that training wasn't designed to actually uncover like what it is that.

00:04:04 Ben Kanspedos

the learner is supposed to do after they take the training and there's no actual behavioral change after that, then yeah, you save 39 hours, but you just wasted 10,000 hours, right?

00:04:16 Ben Kanspedos

So the priority should be the quality, right?

00:04:22 Ben Kanspedos

And you know, for, and of course, speed does matter.

00:04:29 Ben Kanspedos

But you can do it in a way in which you get the speed benefits and you can use a I in a way in which the training is actually effective as well.

00:04:41 RK.Prasad

That's a very, very interesting observation, Ben, when you said that we may be you know, gaining some hours in content generation, but.

00:04:58 RK.Prasad

If we roll out bad courses quickly with A I, then the repercussions are much, much more wide and right, negatively impactful, right?

00:05:18 RK.Prasad

So can you just differentiate between generating content and designing learning?

00:05:30 RK.Prasad

With the backdrop of AI.

00:05:33 Ben Kanspedos

Right?

00:05:34 Ben Kanspedos

I mean, anyone can generate content with AI, but learning has to be designed.

00:05:40 Ben Kanspedos

So generating produces an artifact where design makes a decision about the learner and what they can't do yet.

00:05:51 Ben Kanspedos

and.

00:05:52 Ben Kanspedos

why they can't do those things and what makes it difficult and what needs to be done for them to be able to do those things.

00:06:00 Ben Kanspedos

You know, effective training influences or impacts behavior, right?

00:06:07 Ben Kanspedos

So I'm a big fan of scenario based training because it allows you, it allows the learner to be put in a realistic situation.

00:06:17 Ben Kanspedos

They have to make decisions, they have to think through.

00:06:20 Ben Kanspedos

the consequences.

00:06:23 Ben Kanspedos

And that is what really makes the training effective if they can actually, you know, do the, you know, do some, their behavior is going to be different after they take that training.

00:06:40 Ben Kanspedos

And the thing is, AI is going to do what you tell it to, right?

00:06:46 Ben Kanspedos

So if you tell it to create training.

00:06:49 Ben Kanspedos

and give it source material, it's going to, you know, give you information.

00:06:55 Ben Kanspedos

But if you tell the AI to create like a scenario based training and make sure that training is active and you give the instructions in such a way that it's actually going to create a good training, then it's going to do that.

00:07:10 Ben Kanspedos

So it's just a matter of having the mindset of like how you use the AI and.

00:07:16 Ben Kanspedos

Applying your instructional knowledge to the AI and giving the AI context, right?

00:07:24 Ben Kanspedos

Because it's not, it's not going to, it's not going to know like the business requirements and things like that.

00:07:29 Ben Kanspedos

So it's really just about kind of how you, the input that you put in is going to determine the output that you get.

00:07:42 RK.Prasad

Okay, so what I understood.

00:07:46 RK.Prasad

The difference between generating content is a function of what context and what prompts you.

00:07:55 RK.Prasad

The generate AI will determine the quality of the course.

00:08:01 RK.Prasad

It'll follow your instructions to the T and generate it.

00:08:07 RK.Prasad

Did I get it right?

00:08:09 Ben Kanspedos

Yeah, that's yeah, that's about right.

00:08:13 RK.Prasad

So most people, yeah.

00:08:15 RK.Prasad

Most people are using it to using Gen A I to create those artifacts which go into the course, but they are not looking at the design part of it.

00:08:27 RK.Prasad

In my own personal experience, Ben, people tend to take shortcuts.

00:08:36 RK.Prasad

If there is something very easy to do, they tend to become a little.

00:08:43 RK.Prasad

For a want of a better word, I would say they become a little lazy not checking what the generative AI is throwing out to them.

00:08:54 RK.Prasad

See, I have noticed some instructional designers who, let's say, at the first step, the content comprehension step.

00:09:05 RK.Prasad

So they give all the documents, the raw material, so to speak, different formats, different authors.

00:09:13 RK.Prasad

Out of which the content has to be taken out.

00:09:18 RK.Prasad

Now that is an easy task for Gen AI.

00:09:22 RK.Prasad

It will just spew out a summary.

00:09:26 RK.Prasad

How does the designer make sure that all the important points in the raw content are represented in the summary?

00:09:36 RK.Prasad

Because we know as we have used Gen AI for quite some time.

00:09:41 RK.Prasad

We know sometimes the Gen AI tends to miss certain important points.

00:09:51 RK.Prasad

So how does how do you ensure that it is on track?

00:09:59 Ben Kanspedos

Yeah, I think the first thing is thinking about AI as.

00:10:07 Ben Kanspedos

like a thought partner, how can you work with the AI so it can help expand your ideas, how it can help you think through as opposed to thinking for you.

00:10:23 Ben Kanspedos

And, you know, AI is only as good as the context you give it.

00:10:27 Ben Kanspedos

So there's a lot of platforms where you can upload your source content and it will give you a nice looking training.

00:10:35 Ben Kanspedos

But.

00:10:36 Ben Kanspedos

Did the AI know, does the AI even know like what problem it's really trying to solve?

00:10:42 Ben Kanspedos

Does it understand the business context?

00:10:46 Ben Kanspedos

Does it understand the you know, the audience, the business culture, like where that training sits in the broader training that person is taking?

00:11:02 Ben Kanspedos

You know, I want a platform.

00:11:05 Ben Kanspedos

that expands on instructional designs, not on the instructional designers' inputs.

00:11:12 Ben Kanspedos

And the AI can help extract the information from the human so that it has everything that it needs.

00:11:25 Ben Kanspedos

And I also think there's a opportunity from the evaluation standpoint as well.

00:11:30 Ben Kanspedos

Instead of the AI just generating more, you can use it in a way

00:11:35 Ben Kanspedos

That you have different agents that are doing specific things and there could be one of the evaluation as well.

00:11:44 Ben Kanspedos

So it is actually evaluating the training afterwards, identifying any opportunities, any gaps based on the principles and the rules that you give it.

00:12:01 RK.Prasad

So here you have mentioned.

00:12:04 RK.Prasad

Some A new word here that is agents.

00:12:10 RK.Prasad

Now, most of us in the learning fraternity, we of course know ChatGPT or Cloud or these popular Gen A I tools and we also know certain A I powered authoring tools like for example, I think Articulate.

00:12:33 RK.Prasad

Products do have AI built in now.

00:12:37 RK.Prasad

We have very specialised products like Mid Journey, Beyond.

00:12:42 RK.Prasad

So if you see the learning fraternity, most of their AI usage in using those tools or the Gen AI like Cloud or ChatGPT.

00:12:55 RK.Prasad

So when you're talking about agents, can you please differentiate or like explain what?

00:13:02 RK.Prasad

What is an agent and how is it different from a gen AI platform like ChatGPT?

00:13:09 Ben Kanspedos

Yeah, an agent's basically just AI with some special instructions, right?

00:13:15 Ben Kanspedos

So when you're creating the training, you could have an agent that and you give it the instructions.

00:13:21 Ben Kanspedos

You say, you know, you are, you know, experienced instructional designer.

00:13:28 Ben Kanspedos

I, you know, I want you to be able to.

00:13:31 Ben Kanspedos

challenge me, expand on my ideas, and make sure that the content follows these particular guidelines, right?

00:13:41 Ben Kanspedos

And that's the instructions for that agent.

00:13:44 Ben Kanspedos

And then after you create the training, you have, you know, you have a separate sort of AI instance with different instructions.

00:13:52 Ben Kanspedos

Maybe this is like you're an evaluator, you have these specific criteria, if any of these criteria.

00:14:00 Ben Kanspedos

is below a certain threshold, then send it back, right?

00:14:04 Ben Kanspedos

Don't approve the training until we've hit these specific criteria and guidelines, right?

00:14:10 Ben Kanspedos

So it's really about like setting up a setting up a framework and making sure that your instructional design principles are kind of embedded in this framework and.

00:14:27 Ben Kanspedos

Kind of having a good system as far as how you use the AI and making sure that it has specific instructions and guidelines so that the final output is a certain like level of quality.

00:14:44 RK.Prasad

So an agent is a is a gen AI based tool which has very.

00:14:56 RK.Prasad

Clear specific instructions given by you.

00:15:01 RK.Prasad

Is it a separate piece?

00:15:03 RK.Prasad

Let's say there's ChatGPT.

00:15:06 RK.Prasad

If I open an a chat window and I give all these things, the instructions, the context and all that, is that not an agent?

00:15:21 Ben Kanspedos

Yeah, it is.

00:15:22 Ben Kanspedos

It is an agent.

00:15:22 Ben Kanspedos

It's.

00:15:24 Ben Kanspedos

And it depends on what AI you're using, but you can open up ChatGPT, start chatting, and yeah, it's going to have all the knowledge that ChatGPT has.

00:15:41 Ben Kanspedos

But you can also set it up in a way where when you're chatting with one AI, it has additional instructions that are.

00:15:51 Ben Kanspedos

specifically for the task that you're doing.

00:15:54 Ben Kanspedos

So it's just kind of like setting up a system where when you're doing one task, you're working with an agent that has like a specific skill that you give it.

00:16:06 Ben Kanspedos

And then when the, you know, at different phases in the development, you could have, you could give that content.

00:16:14 Ben Kanspedos

It's like a, it's like a pipeline, right?

00:16:16 Ben Kanspedos

And you can give that content to another agent with

00:16:21 Ben Kanspedos

specific, maybe it's about tone and branding and you can have that agent go through the content and maybe give some feedback.

00:16:29 Ben Kanspedos

And then once you have that, you can know you can send it to another agent and that agent has special instructions maybe from like a compliance and technical standpoint that it that it reviews.

00:16:41 Ben Kanspedos

So it's, you know, depending on what platform you have, it can get kind of complicated.

00:16:46 Ben Kanspedos

But even if you use something very general, it's just a matter of

00:16:51 Ben Kanspedos

kind of separating the work and having having special instructions for the A I that you're talking with at that stage in the development.

00:17:04 RK.Prasad

So it looks, it sounds pretty, I mean, specialised kind of thing making these agents.

00:17:15 RK.Prasad

For instance, I when I write something, I go to a particular project.

00:17:20 RK.Prasad

InChat JPT which I name it as my writing and there I have uploaded some of my writing samples and depending on where I want to publish it, I give the editorial guidelines.

00:17:34 RK.Prasad

So whenever I go and say now I want to write something on this, it follows whatever I have given it in that particular thing.

00:17:43 RK.Prasad

So is that an agent?

00:17:46 Ben Kanspedos

Yeah, I mean an agent is basically.

00:17:51 Ben Kanspedos

Just following special instructions that you write down, right?

00:17:55 Ben Kanspedos

And sometimes it can also involve tools and things, but at the very core, it's just instructions that you're giving the agent like an extra layer of instructions over the, you know, the core knowledge that the A I has.

00:18:12 Ben Kanspedos

Yeah.

00:18:14 RK.Prasad

So do the agents work simultaneously and parallely?

00:18:21 RK.Prasad

If the workflow is defined in such a way.

00:18:25 Ben Kanspedos

Yeah, I mean, if you're using more of just like chat, you're right, like okay, ChatGPT, like like I'll say Claude, right?

00:18:39 Ben Kanspedos

Claude has chat and then there's something called cowork, right?

00:18:43 Ben Kanspedos

And then chat is.

00:18:46 Ben Kanspedos

A little bit more manual, right?

00:18:48 Ben Kanspedos

You probably have to do it, maybe copy or paste into a new session, right?

00:18:51 Ben Kanspedos

Whereas a tool like Cloud Code, you could, you can sort of have the back end kind of automated where, you know, you have these instructions that are already set there.

00:19:03 Ben Kanspedos

So like when you okay answer, you know, when you put in input, it automatically shifts.

00:19:10 Ben Kanspedos

So.

00:19:12 Ben Kanspedos

You, yeah, depending on what you're using, you could have agents running in parallel on the backend.

00:19:18 Ben Kanspedos

But either way, right, even if it's as simple as chatting with, you know, an instructional design specialty AI, and then when you're done taking that and opening up a new session and pasting it in and putting instructions for the evaluator agent to then review that like.

00:19:39 Ben Kanspedos

Right.

00:19:39 Ben Kanspedos

There's different ways that you can do it, but even if you're if you don't have all the fancy tools and even if you're using the very basic A I, you can still apply this concept.

00:19:48 Ben Kanspedos

The only question is kind of how correct streamlined is it going to be?

00:19:54 RK.Prasad

OK, so it sounds very technical.

00:19:59 RK.Prasad

Does a you know an instructional designer without any technical or programming?

00:20:09 RK.Prasad

Coding background.

00:20:11 RK.Prasad

Can he also set up agents, he or she?

00:20:16 Ben Kanspedos

Yeah, absolutely.

00:20:18 Ben Kanspedos

And I think like agents might seem a little foreign and a little complicated, but it can literally be you open up a text document and you.

00:20:38 Ben Kanspedos

Just like if you were telling a person, right, here's your, here's your goal, here's your responsibilities for reviewing this or helping with this.

00:20:49 Ben Kanspedos

It just kind of typing it out, right?

00:20:53 Ben Kanspedos

And then giving that to the AI in addition to any of the training material that you're working on.

00:21:04 Ben Kanspedos

So it's, you know, at the base, at the core, it's really just.

00:21:08 Ben Kanspedos

instructions that you're what you're telling the AI to do.

00:21:12 RK.Prasad

But can you, can you, can you describe a particular project you have personally done in your earlier assignments so that we understand how the whole thing works?

00:21:27 RK.Prasad

Because you have a very unique background of being an instructional designer and also having worked on generative AI for many, many years and you did.

00:21:36 RK.Prasad

Such wonderful work.

00:21:37 RK.Prasad

So a case study, you know, if you can explain the context and how you have gone about doing it, I think it'll be very interesting.

00:21:50 Ben Kanspedos

Yeah, I, you know, one thing that I think that AI can is kind of like underutilized when it comes to learning and development is.

00:22:03 Ben Kanspedos

on the learners side, right?

00:22:06 Ben Kanspedos

Because we're, you know, the industry is really moving to a lot more to self led training.

00:22:16 Ben Kanspedos

And when you move training like kind of out of the classroom or where there's a facilitator, it leaves a big gap because when you're in a classroom,

00:22:29 Ben Kanspedos

a lot of the learning is from the discussion, right?

00:22:32 Ben Kanspedos

Or the facilitator asking questions, and you have to think through it and, you know, articulate your thoughts and role playing and kind of having discussions where you relate what you're learning with your personal experiences, right?

00:22:49 Ben Kanspedos

Like that's where a lot of the learning happens.

00:22:51 Ben Kanspedos

And with, you know, in a self led

00:22:56 Ben Kanspedos

training, you kind of, you know, you miss that, right?

00:22:59 Ben Kanspedos

It's so it's a gap.

00:23:00 Ben Kanspedos

So I created, you know, a training system which allows, which allows the learner to have conversations, like practice.

00:23:12 Ben Kanspedos

So like a use case would be helping managers have difficult discussions with.

00:23:22 Ben Kanspedos

with like their employees, right?

00:23:24 Ben Kanspedos

So like AI plays the role of the employee and the learner plays the role of the manager, and they can practice and have difficult conversations and then get feedback at the end on how you did.

00:23:39 Ben Kanspedos

So it's a cool system, right?

00:23:45 Ben Kanspedos

But giving AI all of the information it needs.

00:23:51 Ben Kanspedos

To have that conversation isn't so straightforward, right?

00:23:55 Ben Kanspedos

If you're doing it for the first time, you know, you have to think about the personality, the specific objectives, the scenario, you know, what information it's going to give when you ask it a question.

00:24:12 Ben Kanspedos

So all of these things need need to be sort of put together ahead of time.

00:24:20 Ben Kanspedos

So

00:24:21 Ben Kanspedos

I created a system which is basically, it was just a basically a bot that had instructions and the instructions were to ask certain questions from the designer who was creating that simulation.

00:24:41 Ben Kanspedos

And it would kind of like do a lot of the heavy lifting and it would kind of give its.

00:24:51 Ben Kanspedos

It would like kind of preview and make kind of take the take the first pass at like giving all of the details that would go into that simulation.

00:25:01 Ben Kanspedos

And so in that case, it was really about like giving the AI instructions on here, here's all the information that is needed.

00:25:10 Ben Kanspedos

Work with the instructional designer to make sure you have everything you need.

00:25:16 Ben Kanspedos

If.

00:25:18 Ben Kanspedos

you know, and analyze it too, say, hey, you know, you wanted me to, you know, you're going to be talking about this topic, but that the AI needs to have this specific information in order for that conversation to go smoothly.

00:25:32 Ben Kanspedos

So like, give me that information, right?

00:25:35 Ben Kanspedos

So it so I created a system which basically used the AI to get all of the relevant information from the instructional designer.

00:25:45 Ben Kanspedos

And then once all of that was gathered and it kind of summarized it for the instructional designer and the instructional designer checked off, then AI took all that information that's needed and that was used as the back end for when the learner was actually having that conversation plus the evaluation that would happen afterwards as well.

00:26:13 RK.Prasad

So.

00:26:15 RK.Prasad

What I understood, Ben, is that in your case, in the case which you described, you have been using A I more like a thinking partner than a substitute for your own thinking, right?

00:26:30 RK.Prasad

And that you can do it through the concept of agents, which means very specific instructions given to that.

00:26:39 RK.Prasad

And there is a conversation going on between.

00:26:43 RK.Prasad

The agent and the human in this case, is that correct?

00:26:48 Ben Kanspedos

Yeah.

00:26:50 Ben Kanspedos

And it is very similar on the learner side, right?

00:26:54 Ben Kanspedos

That you want an experience where the learner has to think through and.

00:27:10 Ben Kanspedos

That the AI's role is to, you know, challenge, push back, reframe, right?

00:27:18 Ben Kanspedos

And the same thing on the instructional designer side as well, right?

00:27:24 Ben Kanspedos

You can use the agent to have the instructions to, you know, evaluate.

00:27:40 Ben Kanspedos

Poke holes, push back on anything that's where there's opportunities, right?

00:27:48 Ben Kanspedos

So if you, it's kind of like a mental shift, right?

00:27:53 Ben Kanspedos

It's helping AI think with you, not for you, right?

00:27:59 Ben Kanspedos

Like in school, for like in class, I know there's like, there's kind of like debates, right?

00:28:07 Ben Kanspedos

Like.

00:28:08 Ben Kanspedos

AI is bad because people don't think they just, they just ask AI for the answer and they put it in there, right?

00:28:17 Ben Kanspedos

And if the person using the AI is using it in that way, where it's just asking for answer, then yeah, that is sort of bypassing the learning.

00:28:29 Ben Kanspedos

But if the AI, if it's more of.

00:28:35 Ben Kanspedos

a situation where you're the AI is specifically designed to make sure that you understand it better and that you're constantly improving, then AI goes from a thing where it's reducing the training to a thing where it's really helping you learn faster and helping uncover things that you wouldn't have thought to, you know, that you wouldn't have thought to ask or that you wouldn't have known was, you know, an opportunity.

00:29:05 Ben Kanspedos

So it's the same thing with instructional.

00:29:06 Ben Kanspedos

You know, when you're doing the instructional design, you can use it in a way in which it's programmed to help you become a better learner as you go and as you use it to develop the training.

00:29:23 RK.Prasad

Well, that that is a that's a totally different kind of observation you made.

00:29:31 RK.Prasad

So if we use A I agentic A I in the way you described, we will not only have very high quality courses, but it will also make learning professionals much more competent and much more qualified and better.

00:29:49 RK.Prasad

So it is like something like an instructional designer working with a very superior kind of instructional designer together.

00:29:59 RK.Prasad

So in the process.

00:30:01 RK.Prasad

I will learn a lot.

00:30:04 RK.Prasad

My thinking will improve, it will become more comprehensive and probably expand my horizons when I design a course.

00:30:13 RK.Prasad

So it will be a continual learning.

00:30:16 RK.Prasad

It will not be just a tool to develop courses, but also to develop oneself.

00:30:25 RK.Prasad

Yeah, I think that is a very fascinating.

00:30:30 RK.Prasad

scenario where the tool helps you become better, right?

00:30:38 Ben Kanspedos

Yeah.

00:30:38 Ben Kanspedos

And it's, it's, you know, it's a subtle shift, but it's, yeah, it makes all the difference, right?

00:30:45 Ben Kanspedos

It's a difference between saying, all right, here's some content and, you know, create a training with this and it will take the content and do it and versus.

00:30:59 Ben Kanspedos

I'm creating this training.

00:31:00 Ben Kanspedos

Here's the source content.

00:31:02 Ben Kanspedos

Here's what we're really trying to allow the learner to do, and here's what we need to accomplish.

00:31:10 Ben Kanspedos

What's the best way of doing that?

00:31:11 Ben Kanspedos

Right?

00:31:12 Ben Kanspedos

Like Let's brainstorm and let's think through ideas together on like what the best way of like helping that learner accomplish those things in this training.

00:31:19 Ben Kanspedos

It's sort of like just asking it in that way.

00:31:23 Ben Kanspedos

Because like I said, it's going to kind of do what you ask it to do, right?

00:31:27 Ben Kanspedos

So it's just sort of having that mental shift in as far as like how you, it's more of like thinking about like using AI as, you know, a way of the way that you would talk with a person, right?

00:31:45 Ben Kanspedos

Versus like a Google where you're just like wanting an answer.

00:31:51 RK.Prasad

Thanks.

00:31:53 RK.Prasad

So this is a very, very fascinating and very interesting times for all professionals, especially the learning professionals.

00:32:01 RK.Prasad

I would like to change gears a little here, Ben, and go a little, take a step back into training needs before we even start thinking of designing courses.

00:32:16 RK.Prasad

You know your performance gaps, your business results.

00:32:20 RK.Prasad

Translating that into training leads and then into measurable learning objectives.

00:32:26 RK.Prasad

We know the routine there now.

00:32:32 RK.Prasad

A I is in all functions of a business.

00:32:36 RK.Prasad

It is in sales, marketing, production, quality everywhere.

00:32:45 RK.Prasad

So will there be a time when?

00:32:49 RK.Prasad

Very personalised training needs can be identified by A I agents working in those functions and feed the LDAI so that every single person based on where he is failing or succeeding.

00:33:09 RK.Prasad

A course it can be put together of course with the intervention of human beings.

00:33:14 RK.Prasad

Do you foresee such a situation where just in time?

00:33:18 RK.Prasad

Just enough very individual learning can be designed with AI.

00:33:26 Ben Kanspedos

Yeah, I think it definitely can.

00:33:30 Ben Kanspedos

To some extent, it's the job that you're doing, right.

00:33:35 RK.Prasad

Correct.

00:33:36 Ben Kanspedos

It's a little bit easier in certain jobs, right?

00:33:38 Ben Kanspedos

Like, so for example, if you're in sales and you're talking with customers.

00:33:45 Ben Kanspedos

Then most businesses already sort of record and transcribe, right?

00:33:51 Ben Kanspedos

So the A I would be able to analyze based off of the instructions and see where, like, what was the biggest opportunity in that last call where, right, you know, you didn't.

00:34:13 Ben Kanspedos

you know, acknowledge the customer's needs or, you know, like you didn't have it, you jumped right into selling without really like listening for straight.

00:34:21 Ben Kanspedos

So, and then also, if there is a system in which the AI has a continual memory of that individual and it are it and it starts to form kind of understand where the.

00:34:40 Ben Kanspedos

strengths and where the opportunities are for that learner, then it's, you know, the technology is already there, right?

00:34:48 Ben Kanspedos

Like we've had this technology for a while to be able to do.

00:34:50 Ben Kanspedos

It's just about kind of putting them together and having a different kind of mindset on how you apply it.

00:34:58 Ben Kanspedos

But yeah, I mean, this is what I, you know, I've been doing this in different ways as far as making the A I like making it more.

00:35:09 Ben Kanspedos

personalized, making it more adaptive.

00:35:12 Ben Kanspedos

And, you know, it's all about the context, right?

00:35:15 Ben Kanspedos

So if it understands like the business biggest opportunities, if it can like analyze the different transcripts from different employees, and like what you're saying here, if it has context about that particular user, the more context you give it, the more targeted and better kind of

00:35:38 Ben Kanspedos

Instructions that it that it can give.

00:35:43 RK.Prasad

So is it also possible that this kind of, let's say I'm a salesperson and I'm having a call with a prospective customer.

00:35:54 RK.Prasad

Now what you're saying is that script, that recording will be analysed by.

00:35:59 RK.Prasad

A I and it can point out where I have missed some opportunities or I didn't listen properly or I didn't understand properly.

00:36:07 RK.Prasad

I could have done this.

00:36:10 RK.Prasad

That is more like, you know, after the fact, a postmortem.

00:36:14 RK.Prasad

Will it be possible that this kind of feedback comes real time?

00:36:24 Ben Kanspedos

You know, from my experience?

00:36:28 Ben Kanspedos

This sort of concept, right, of like real time help has always sounded good, but implementing it in a way that it's actually like helpful and not just in the way has been difficult.

00:36:44 Ben Kanspedos

However, I think we are getting to the point.

00:36:50 Ben Kanspedos

Maybe we're there now, right where.

00:36:55 Ben Kanspedos

You know, generative A I can be, you know, it's not just simple like an algorithm, right?

00:37:01 Ben Kanspedos

We're like, well, you said these keywords, so let's give you this article, right?

00:37:04 Ben Kanspedos

Like that's where it's been like, yeah, it sounds good, but yeah, it's no one likes it, right?

00:37:12 Ben Kanspedos

So I haven't personally seen like a really good implementation.

00:37:17 Ben Kanspedos

Where like AI is sort of like listening and then like, oh, you forgot to say this, right?

00:37:22 Ben Kanspedos

I think that's, but I think we're getting close, right?

00:37:24 Ben Kanspedos

And there, I'm sure someone out there is doing it effectively, but I think like it's always been a thing where like we really weren't ready for it to like actually do it well.

00:37:35 Ben Kanspedos

But I think we're definitely getting closer to that now, right?

00:37:43 RK.Prasad

So even.

00:37:44 RK.Prasad

The same scenario if you port it to, let's say, a counselling session between a manager, a performance review between a manager and an employee, or some kind of interaction of conflict resolution, you know, where real time AI help can probably give better solutions.

00:38:08 Ben Kanspedos

Yeah, you know most.

00:38:13 Ben Kanspedos

We're in a pretty virtual world, right?

00:38:15 Ben Kanspedos

And I think one really underutilized way of using AI is to, is like the conversation, right?

00:38:26 Ben Kanspedos

Like I did, you know, I've been doing some workshops to teach people Cloud Code, right?

00:38:33 Ben Kanspedos

And it's on Zoom and it's recording.

00:38:36 Ben Kanspedos

And after that, I just take the transcripts.

00:38:42 Ben Kanspedos

An AI has, it's like, you could say it's an agent, right?

00:38:45 Ben Kanspedos

It's instructions are to review, give feedback, understand where in the conversation that maybe you missed an opportunity or like what you did well, and then giving you that feedback.

00:39:00 Ben Kanspedos

And then almost for sure, you're going to see things that's like, oh yeah, I didn't even think about that.

00:39:06 Ben Kanspedos

I totally could have done that differently, right?

00:39:12 Ben Kanspedos

And that concept can be applied, you know, whether you're doing like a sales presentation, whether you're counseling somebody.

00:39:19 Ben Kanspedos

Of course, there's there's some privacy things you want to like make sure, you know, you're doing it in the right way.

00:39:25 Ben Kanspedos

But yeah, it's such a, you know, it's such an easy and underutilized thing, I think, to be able to, you know, go back and it could be like a....

00:39:37 Ben Kanspedos

you know, an employee talking with taking a taking a support call, right?

00:39:41 Ben Kanspedos

Like there's so many ways of doing that.

00:39:43 Ben Kanspedos

And then as you were saying, can that be real time?

00:39:46 Ben Kanspedos

I mean, yeah, it's like afterwards, you give the whole transcript, it takes a little while and gives you feedback.

00:39:53 Ben Kanspedos

So it needs to be pretty good because because it's going to be like, it needs to be reading and analyzing in real time and being able to like proactively provide the right information in the right way.

00:40:07 Ben Kanspedos

So like I said, I haven't, I haven't personally seen a good effective implementation of that, but it's definitely getting close and it's definitely it it's definitely except for the speed thing.

00:40:23 Ben Kanspedos

I think that's like the probably like the biggest obstacle to get through right now.

00:40:29 Ben Kanspedos

Because if you're if you're talking to.

00:40:31 Ben Kanspedos

With somebody and you're getting live like it has to be like very fast, right?

00:40:35 Ben Kanspedos

So I think that's probably like.

00:40:37 RK.Prasad

So you get distracted or it will become very stilted.

00:40:41 RK.Prasad

You know, I can't say, Ben, just a minute, let me read what my AI tutor is telling you.

00:40:48 Ben Kanspedos

Yeah.

00:40:48 Ben Kanspedos

And then there's also like the danger maybe of just becoming too reliant on it and then you're not really having a real conversation, right?

00:40:59 Ben Kanspedos

So that that's why I, I mean, I've personally.

00:41:02 Ben Kanspedos

been a little bit has like skeptical about this like real time.

00:41:05 Ben Kanspedos

We can help you in real time, but I'm, I don't know.

00:41:08 Ben Kanspedos

I'll believe it when I see it.

00:41:10 Ben Kanspedos

You know, I know we've been trying to do it for years even before AI.

00:41:15 RK.Prasad

Right?

00:41:16 RK.Prasad

Actually, Ben, I just recall my own personal experience when I was in pharmaceutical sales.

00:41:27 RK.Prasad

The method here in India was that.

00:41:30 RK.Prasad

When you first go into the field, you finish your classroom training and you start visiting the doctors.

00:41:36 RK.Prasad

A senior rep will come along with you, just observe.

00:41:40 RK.Prasad

And when you step out of the chambers, he immediately gives you feedback.

00:41:47 RK.Prasad

You understand so that now AI can do at any scale and in a much better way, right?

00:41:54 Ben Kanspedos

Yeah, I mean, right.

00:41:57 Ben Kanspedos

It really level, you know, it levels the playing field, right?

00:42:01 Ben Kanspedos

Because, right.

00:42:04 Ben Kanspedos

Most, you know, people who get who have like a dedicated trainer guide that can help them.

00:42:14 Ben Kanspedos

You know, not everyone has that, right?

00:42:17 Ben Kanspedos

So it's kind of.

00:42:21 Ben Kanspedos

An imbalance where like more people who have these people who have someone that can help them kind of have an advantage, right?

00:42:30 Ben Kanspedos

But if you, in a lot of ways, AI can be a better tutor than a human, right?

00:42:43 Ben Kanspedos

Because it can.

00:42:47 Ben Kanspedos

give instructions in a way where it's not just giving you the answer, it's it's asking questions in a way that you make the conclusions yourself, right?

00:42:58 Ben Kanspedos

And, you know, you get the, you get the feedback immediately and it's able to like explain it in a way like that's very difficult, even for like a really good human tutor to do consistently, right?

00:43:12 Ben Kanspedos

So I think as you know.

00:43:17 Ben Kanspedos

technology becomes more available for everybody and it advances.

00:43:22 Ben Kanspedos

I think it's really going to help, you know, it's going to have the opportunity of helping everybody as opposed to, you know, just people who kind of are privileged enough to have an actual human tutor kind of helping them along the way.

00:43:45 RK.Prasad

Yeah.

00:43:46 RK.Prasad

Yeah, that is something we can look forward to.

00:43:50 RK.Prasad

As you rightly said, its already there in some form or the other.

00:43:55 RK.Prasad

Have you?

00:43:56 RK.Prasad

I am sure you this organization, Khans Academy, Khan Academy.

00:44:03 RK.Prasad

Have you heard of this?

00:44:05 RK.Prasad

Yes, Its a free platform which teaches school and college kids various subjects.

00:44:16 RK.Prasad

Now I read that they have tied up with ChatGPT and they have a tutor inside, right?

00:44:26 RK.Prasad

So I actually paid the fee and I tried to see it.

00:44:34 RK.Prasad

So that tutor, I just can't recall the name.

00:44:37 RK.Prasad

There is some Spanish name which is given to the tutor there.

00:44:42 RK.Prasad

Is exactly what you are you just mentioned.

00:44:45 RK.Prasad

It does that.

00:44:46 RK.Prasad

If a student asks a question, it will answer in a very indirect way and asks a question back.

00:44:55 RK.Prasad

So it makes sure the student is engaged, right?

00:45:00 RK.Prasad

So it has already gone into the system.

00:45:02 RK.Prasad

But my last question to you, Ben, actually it is A and B.

00:45:10 RK.Prasad

How do you foresee L&D in the next 5 years?

00:45:16 RK.Prasad

That's my question.

00:45:18 RK.Prasad

A and B is how do learning and development professionals prepare themselves for it?

00:45:27 Ben Kanspedos

Yeah, good question.

00:45:28 Ben Kanspedos

And yeah, it is funny you mentioned the Khan Academy because I saw when I.

00:45:36 Ben Kanspedos

Maybe 3 or 4 years ago when I was first getting into this and figuring out like how can we use AI?

00:45:45 Ben Kanspedos

I saw a TED Talk by the guy that runs Khan Academy and he was talking about these exact concepts, right?

00:45:54 Ben Kanspedos

And it was really about about you know, they had the right.

00:46:05 Ben Kanspedos

intention and like they designed the AI in a way purposely.

00:46:09 Ben Kanspedos

They designed in the AI in a way because they understood like the learning principles, right?

00:46:17 Ben Kanspedos

So they designed the AI to not just give answers, but ask the right questions.

00:46:24 Ben Kanspedos

So the learner comes to those conclusions on their own right?

00:46:30 Ben Kanspedos

And right if.

00:46:31 Ben Kanspedos

It's sort of like the push versus pull, right?

00:46:34 Ben Kanspedos

Like if you're just pushing information, you're not going to learn as much.

00:46:37 Ben Kanspedos

But if you do it in a way in which the person is like engaged and like pulling the information, then yeah, that's that's where the learning's going to happen.

00:46:48 Ben Kanspedos

So I think it's just about like the architects who are putting these things together, if they're if they have that instructional design, you know, mindset, then they're going to create

00:47:03 Ben Kanspedos

the experience so that learning happens, right?

00:47:07 Ben Kanspedos

So I think it's important that like it's not just about the technology, right?

00:47:13 Ben Kanspedos

It's like about the principles that go into it.

00:47:20 Ben Kanspedos

So yeah, I don't know if I think that might have answered to your first question, right?

00:47:25 Ben Kanspedos

What was Remind me again.

00:47:27 Ben Kanspedos

What was that the second part of your question?

00:47:30 RK.Prasad

The second part is that how people like us learning professionals prepare ourselves, train ourselves to use these wonderful you know, developments for the benefit of our own careers as well as to those whom we are serving, the learners.

00:47:53 RK.Prasad

How do you think that we need to go ahead and?

00:47:58 RK.Prasad

Learn how to do these well.

00:48:02 Ben Kanspedos

Yeah, I think it goes back to just understanding that A I is just a tool, right?

00:48:10 Ben Kanspedos

And it's a very powerful tool.

00:48:12 Ben Kanspedos

And it's only going to be as good as the person using like how you use that tool.

00:48:19 Ben Kanspedos

Like if you have, you can have a very.

00:48:23 Ben Kanspedos

powerful tool, but if you don't use it right, then it's not, you know, going to have, it's not going to be like very helpful.

00:48:29 Ben Kanspedos

So not just relying on AI as if it's just going to magically do what you want it to do, but really having having a like a learning mindset and always, you know,

00:48:53 Ben Kanspedos

I can say, I can say from my standpoint, the more I use AI, the more it started to make sense and the more that I could kind of think in like a more systematic way.

00:49:10 Ben Kanspedos

And the more I use it, I just like anything you learn, right, your brain gets, the more exercise your brain gets in a certain area, the better, you know.

00:49:22 Ben Kanspedos

It kind of tunes to that kind of skill set.

00:49:26 Ben Kanspedos

And it's a similar thing, right?

00:49:28 Ben Kanspedos

Like you can use the technology, but there's no, there's no shortcut for experience, right?

00:49:34 Ben Kanspedos

And so to some extent, I think it's just being curious, trying new things, be experimental and things will start to click and you'll just get better and better at it, right?

00:49:50 Ben Kanspedos

And.

00:49:52 Ben Kanspedos

The great thing about AI is that it not only can perform for you, it can also, like we were saying, it can be a thought partner.

00:50:01 Ben Kanspedos

You can just ask it, hey, I'm trying to do this one thing, what would you suggest?

00:50:05 Ben Kanspedos

And it can kind of give you these things.

00:50:07 Ben Kanspedos

And then in that process, you're kind of learning.

00:50:11 Ben Kanspedos

You're like, oh yeah, that's that's that's a good idea, right?

00:50:14 Ben Kanspedos

And then you can kind of guide it, say like yeah, I like this idea, but how do we make it more.

00:50:21 Ben Kanspedos

like realistic, right?

00:50:22 Ben Kanspedos

And it's kind of like going back and forth.

00:50:25 Ben Kanspedos

So it's really like the technology is there.

00:50:29 Ben Kanspedos

I think the biggest opportunity is just having the kind of like mindset and the shift as far as how you use that.

00:50:37 Ben Kanspedos

And it really goes back to using AI to think with you, not for you, right?

00:50:44 Ben Kanspedos

So I think it's just about using it and kind of having that mindset and.

00:50:48 Ben Kanspedos

Not using it to replace your thoughts, but to expand on your thoughts.

00:50:54 RK.Prasad

Right.

00:50:55 RK.Prasad

I think that is a very, very useful advice that you have given in podcast was, if I may summarize, your message is that use AI as your thinking partner, not as a substitute for your own thinking.

00:51:13 RK.Prasad

Be curious, learn and more you learn, more you interact.

00:51:18 RK.Prasad

With your partner, you become better and better.

00:51:21 RK.Prasad

That is one clear message which is coming out.

00:51:24 RK.Prasad

Don't make the mistake of using it for only speed.

00:51:31 RK.Prasad

Then you will generate or create courses which are of low quality very fast.

00:51:38 RK.Prasad

So I that resonated very well with me, I am sure with our audience also.

00:51:43 RK.Prasad

Ben, that has been very wonderful.

00:51:47 RK.Prasad

Thank you very much for your time and for sharing these very, I would say, incisive observations.

00:51:57 Ben Kanspedos

You're welcome.

00:51:58 RK.Prasad

Yeah, that's really take back a lot to the.

00:52:01 RK.Prasad

Thank you very much.

00:52:04 Ben Kanspedos

Yeah, this is really what like I've been thinking about for a long time.

00:52:08 Ben Kanspedos

So it's really good to kind of talk with you RK and having these sort of.

00:52:15 Ben Kanspedos

talking at this, at these topics, right?

00:52:19 Ben Kanspedos

Because I think we're kind of at a cusp where we're not really there yet, but we're starting to.

00:52:27 Ben Kanspedos

So I think you're really asking the right questions.

00:52:29 RK.Prasad

Yeah, absolutely, absolutely.

00:52:32 RK.Prasad

Yeah.

00:52:34 Ben Kanspedos

So thank you.

00:52:35 Ben Kanspedos

It was great talking.

00:52:35 RK.Prasad

Yeah, it was very surprising that I was independently thinking about certain things, and I found that you were also thinking the same things.

00:52:44 RK.Prasad

Halfway apart, the whole globe, you are on the other side.

00:52:48 RK.Prasad

But that was a very pleasant surprise, and I hope that we will be in touch and continue to work on something together.

00:52:57 Ben Kanspedos

Yeah, I think we are, like I said, I think we are kind of at the cusp of really like the next level of how we can use AI in learning and development.

00:53:10 Ben Kanspedos

and.

00:53:11 Ben Kanspedos

Yeah, it's great to have someone on the same same wavelength talking about these things.

00:53:15 RK.Prasad

Pleasure is all mine, Ben.

00:53:17 RK.Prasad

You have a great day.

00:53:18 RK.Prasad

Thank you very much.

00:53:20 Ben Kanspedos

You too.

00:53:20 Ben Kanspedos

Okay, thank you.