Performance Support

Contextual Help



Why Generative AI Isn’t Suitable for Most L&D Tasks, Especially Corporate IT Training

Since OpenAI’s ChatGPT meteoric rise in November of 2022, it seems to have leveled out as the reality of its uses (and lack of uses) become evident. It has also become apparent that ChatGPT isn’t the only player in the generative AI market.

I’m sure many industries were excited by the recent power of generative AI, but L&D was exceptionally excited. It’s going to replace jobs! No, wait, AI won’t replace jobs, but someone using AI will replace jobs! No, wait, it’s not going to replace any jobs.

You can use it to create courses and write your entire script… The list of things it supposedly can do goes on and on. The problem is that it has so many unrepairable flaws that it won’t replace any jobs. It’s not even likely that someone using AI will replace the job of someone not using it.

Of course, it does have a few uses, but it’s not as powerful as some make it out to be. In fact, generative AI isn’t suitable for most L&D tasks. That goes doubly so for technical training or company IT training.

Generative AI won’t take your L&D job, and neither will someone using AI.

Nick Leffler

It’s useful, of course, and it has its place in everything to help humans perform better. But it’s really really really dumb. It’s overly confident, full of fake information, and constantly spouts non-truths, and it just makes way too much up to be a significant threat to people.

ChatGPT's answer to my question about using learning styles to guide my training design.
Which one is it? Should I or shouldn’t I use learning styles to guide my training ChatGPT? (hint: you shouldn’t, and ChatGPT isn’t nearly as definitive as it should given the research)

AI is revolutionizing how we live and work, from self-driving cars to voice assistants, but it hasn’t revolutionized L&D nearly as much as we’d like. I’d go as far as saying that if generative AI has replaced your job or any significant portion of it, your job wasn’t that important to begin with.

I’m not saying someone’s job isn’t important; I’m just saying that generative AI likely hasn’t replaced anyone’s job or even reduced it significantly. That’s because it drones on and repeats itself so much as to be of little use for real learning.

But that hasn’t stopped L&D from trying to shoehorn generative AI into any place it can! It’s primarily vendors, though, with their fancy AI-driven LMSs and dev tools with AI built in. Great, more useless content that’s all a bunch of fluff…

Generative AI is great at that, and there’s nothing better than creating overwhelming and useless training.

Generative AI is confident, knows everything (according to it) and will never let you down by not providing you an answer even if it’s wrong.

In case you’re unaware, generative AI is a subset of AI that focuses on creating new content based on existing data, such as images, videos, or text. It uses complex algorithms and machine-learning techniques to generate original content miming human-like creativity. Sounds impressive, right?

Not really. It’s like an over-glorified auto-complete for your text messages. So, before you jump on the generative AI bandwagon, there are certain considerations, especially regarding corporate IT training.

The use of generative AI for company IT training seems enticing at first. Imagine a system that can generate training materials on the fly tailored to employees’ specific needs. It could save time and resources, eliminating the need for human trainers or instructional designers.

However, as enticing as it may sound, generative AI falls flat in almost every crucial area regarding L&D tasks. It especially falls flat at corporate IT training, both creating content and simply writing it.

It can come in handy in a few places but is limited. It’s limited and requires so much oversight that it ultimately saves you very little time if any at all. Let’s look at generative AI for L&D and IT training.

The Potential of Generative AI in L&D

It’s not all negative for generative AI. There are some excellent uses for it, and for creativity, it can get your brain going or provide you with a basis to go off of and edit heavily.

These are a few ways we’ve found feasible to use generative AI for workplace training, including technical training. One essential thing to remember is that no proprietary or private information should ever be put into any AI tool.

In general terms, it can be decent to explain a general goal you’re trying to achieve, and then it can provide you with a decent place to start. Here are some ways it could be helpful:

  • It can create a scenario outline that you can then customize.
  • Create fictitious characters and personas to use in a course.
  • Have it generate test questions if your information is general in nature.
  • Summarize long content.
  • Provide a conversation transcript with a subject matter expert and summarize that information.

Generative AI is great at making stuff up completely. That means it’s excellent for creative work, stories, and information that’s not factual. It will make information up and is confident it’s right until you tell it it’s wrong, and then it will still go back to thinking it’s right (it’s forgetful).

But for that creative work such as those scenarios or creating fake characters, branching, and scenarios, it’s great. Our entire CRM showdown example microlearning course in our portfolio began as AI content with two fake CRMs and client needs.

Generative AI can help get you started at simple tasks that requires no accuracy, succinctness, or completeness.

It’s creative and fun, and it lets you learn reasoning and decipher client needs, but none of it’s real. It’s just for fun but can strengthen decision-making skills. So, there are uses for it; it’s just not as useful as some make it out to be.

So, let’s take a look at the limitations of generative AI.

Limitations of Generative AI in Company IT Training

There are many more limitations of generative AI in L&D than there are uses. That doesn’t mean it’s useless altogether. It means you must use it appropriately and not expect too much from it.

Some are so taken by AI that you’d think it saved them from death. But in every presentation, I’ve seen that the examples of its output, even with the best prompt possible, are mediocre. Not to mention its lack of knowledge and ability to put nearly anything factual into writing.

Let’s take a look at some of the more important limitations to take into account. These limitations apply to most L&D in general, but I’ll be speaking to them through the viewport of IT training.

Lack of Domain Expertise

One major limitation of generative AI in company IT training is its lack of domain expertise. While it can generate content based on existing data, it does not possess the deep understanding and knowledge that subject matter experts have.

It’s also prone to a lot of false information that a true expert would be able to decipher. Heck, sometimes it can’t even decipher information that someone with no expertise would be able to.

An excellent example of this is when I asked ChatGPT if it could give me the steps to create a meeting now in Microsoft Teams. It wanted me to click the New meeting button in the Calendar tab. I asked it how to create a meeting now, which means it should have been able to tell me to click the Meet now button rather than New meeting.

That’s just one example of it not having any expertise in anything, and there’s a good chance it can make stuff up completely. I wouldn’t trust it for anything that needs to be factual or correct.

Generative AI makes up a lot of information (even citations) so it can’t be relied upon as an expert in anything.

It also doesn’t know your company’s policies or processes. Those two facts are enough to rule it out as any real contender to help create corporate IT training.

However, there will be many AI tools that allow for company-specific information. Unfortunately, we’ll have to wait and see if those can provide intelligent processes. Even if ChatGPT is given accurate information, it’s still prone to mistakes.

In corporate IT training, it is crucial to have subject matter experts and instructional designers who possess extensive domain expertise and can provide real-world examples and insights.

Experts can answer questions, clarify doubts, and guide employees through practical exercises. Generative AI simply cannot replicate this level of expertise and will never be an expert on any technical topic. It’s more likely to grab false information and mislead someone who doesn’t know better.

Inability to Address Individual Learning Needs

Another limitation is the inability of generative AI to address individual learning needs. While it can provide personalized recommendations based on performance data, it lacks the emotional intelligence and empathy that humans possess.

Humans (experts and instructional designers) know the emotions and motivations of employees and can build interesting and valuable training for employees’ specific jobs. Good training relies on relevant training, and generative AI will likely always lack the skill to create the right stuff.

There are a lot of subtleties in training that experts and AI might gloss over. Instructional designers can’t be replaced because they can simplify content and ask essential questions of experts to build the best training possible for employees.

Limited Interactivity and Personalization

Generative AI may be able to generate generic training materials, but it lacks interactivity and personalization. Interactive learning experiences, such as engaging and relevant scenarios or simulations, are crucial in company IT training to ensure employees gain practical skills.

While generative AI can create scenarios, they’ll never be as good as what an expert could come up with. That’s because experts know the actual working conditions and likely have real scenarios they’ve experienced.

Scenarios that generative AI will never be as realistic or good as someone who has actually experienced an interesting situation on the job.

Turning a real story from a SME into a scenario in a course is powerful and effective for helping employees learn. Generative AI will never be able to do that. It can merely come up with generic and not very imaginative scenarios that will never have the same impact as something real.

Not to mention, I’d love to see AI build a realistic software simulation that weaves in engaging scenarios. Good technical training isn’t simply tasks or steps, it’s how to apply them in the flow of work.

Those are some of the major points of generative AI’s limitations for L&D and company IT training. There are also some ethical considerations too.

Ethical Considerations

Aside from the limitations, there are also ethical considerations when using generative AI in company IT training. The use of AI raises concerns about data privacy and security. Companies must ensure that employee data is protected when using generative AI for training purposes.

But it’s not just about employee data; it’s about proprietary company information that can quickly become part of the AI cloud mass. Just like in Vegas, what happens in AI stays in AI forever…

There is generative AI that supposedly doesn’t learn from your company data, but do you trust that? There’s not a good way to keep tabs on that, and are companies like OpenAI able to effectively and fully contain content?

There’s also the fact that generative AI may lead to a devaluation of human expertise. Just because AI can do it doesn’t mean it can do it well or provide any value.

I can give you a tattoo, really.

A dog giving a guy a tattoo.

That’s kind of like how it is when companies put AI capabilities in their tools and say it can create all this content for you. Yes, but is it good, useful, or even accurate content?

Tracy Morgan saying no, no, no.

I don’t think so!

It’s essential to recognize the value people bring and not rely on AI for many training needs if any at all. Not to mention, it can be challenging to validate the information from AI because it’s so lenient with the truth.

The Importance of Human Involvement in IT Training

There will never be a replacement for people in IT training. From the expertise to the design of training, people can’t be replaced. There can only be a small amount of augmentation by generative AI, but that is even limited.

Let’s look at some ways generative AI cannot replace humans and never will be.

Expertise and Experience

People possess expertise and experience that generative AI cannot replicate. They have in-depth knowledge of their respective domains and can provide real-world examples and insights that enhance the learning experience.

Computers can’t replace the experts behind company technology and processes. That is doubly so for instructional designers. Instructional designers are experts in people and how adults learn. They can take too much information from a SME and boil it down while chunking it into something digestible for employees and won’t leave them with train brain.

People are the real experts and computers will never replace that.

With training, nothing is important if everything’s important, and an instructional designer can cut it down to just the essential facts; generative AI can’t do that. It’s verbose and great at writing a lot of fluff for no reason.

An instructional designer paired with a real expert can create amazing and effective training that draws from real expertise, experiences, and practical tips and tricks with realistic scenarios.

This practical knowledge is invaluable in company IT training, as it helps employees navigate real-world challenges effectively and apply what they learn directly on the job.

Emotional Intelligence and Empathy

People possess emotional intelligence and empathy, which are crucial in creating effective learning environments. They can understand the needs of employees and how to accomplish tasks best. Generative AI simply can’t understand that and doesn’t understand how people work.

In company IT training, where employees may feel overwhelmed or intimidated by technical concepts, having content built in a way for people makes a significant difference. Things made by people for people motivate employees, boost their confidence, and foster a culture of continuous learning.

Generative AI can’t apply emotional intelligence or empathy when creating training. An instructional designer can understand how people learn and how to help them learn specific content in the best way possible.

Finding the Right Balance: Combining AI and Human Touch in L&D

There are limitations in generative AI as well as potential uses. It’s undeniable that human involvement in IT training is essential, though. Sometimes, it’s possible to find a way to use generative AI without compromising the business or employees.

There are some excellent ways to combine AI with the needs of people and the need for a human touch in company IT training. Using AI for certain tasks as a starting point is possible but not to complete the work. AI can be a starting point, but that doesn’t mean it will save you 50% of your time.

We’ve found that any task AI can help with, it can save anywhere from 5-25%, but occasionally, it even takes longer. It’s all about finding the right balance for your workflow to create a process of using generative AI when it makes sense.

There’s a balance when using generative AI to help you work. As long as it’s a sidekick that’s kind of bad at what it does and can’t be trusted.

We’ve found a good balance between humans and computers in certain circumstances. Generative AI will likely never effectively eliminate many jobs, but it could positively augment some.

For things that don’t require precise processes or facts, generative AI can provide value. Just don’t count on it providing something you can use as is. Anything output from a computer needs to be heavily reviewed and likely heavily rewritten to make it more accurate, less repetitive, and more helpful.

Wrap Up

Generative AI has the potential to help with various tasks in learning and development. However, when it comes to corporate IT training, significant limitations mean it will never be an ideal solution.

The lack of domain expertise, inability to address individual learning needs effectively, limited interactivity and personalization, and potential technical inaccuracies make generative AI unsuitable for most L&D tasks in company IT training.

People bring way too much value both in expertise and instructional designers. They simply can’t be replaced with their expertise, experience, emotional intelligence, and empathy and cannot be replicated by AI.

There’s potential in generative AI as an assistant but not enough to ever replace any significant job.

By finding the right balance between AI and human involvement in L&D, employees can use generative AI in some ways as long as it’s done carefully, it’s just not going to be as revolutionary as some make it out to be.

With the right uses and balance, creating effective training programs can combine the benefits of technology with a human touch. For the training we create for technical topics, there has yet to be any time we’ve applied generative AI. But AI might assist in a scenario where the subject matter expert simply can’t come up with one.

If you’d like to discuss your next company IT training initiative, we’d love to discuss it with you and see how we can help it succeed. Schedule a free consultation for us to talk; just don’t count on us trying to pass off generative AI work as our own. You’ll likely be able to tell the difference.

Leave a Comment