Click here to watch the full discussion
Most conversations about AI in learning are still focused on the wrong thing.
Everybody wants to know how to generate courses faster, create slides automatically, or turn documents into eLearning. Useful? Sure. But that’s probably not the biggest long-term shift AI is creating inside corporate learning.
There’s a lot of noise right now around AI in learning. Every conference has it. Every vendor has it. Every product suddenly has “AI-powered” attached to the logo somewhere. But underneath all the hype, there’s a much more interesting conversation happening, one that’s less about generating another course faster and more about fundamentally rethinking how learning content works.
That was the focus of a recent LinkedIn Live discussion hosted by David Mantica featuring Robert Gadd. Between the two of us, there’s probably 50+ years of collective scar tissue in the corporate learning industry. We’ve seen the rise of LMS platforms, mobile learning, SCORM, xAPI, gamification, microlearning, learning experience platforms, and now AI. Some trends faded. Some became table stakes.
AI feels different. Not because it’s magical, and not because it replaces expertise, but because it has the potential to fundamentally change how organizations understand, organize, personalize, and deliver learning.
Most AI Discussions Are Focused on the Wrong Problem
Right now, most AI conversations in learning revolve around content generation. “How fast can I create a course?” “How quickly can I generate slides?” “How do I turn a PDF into eLearning?” Those are useful conversations, but they’re also somewhat obvious.
The more interesting challenge is this: most organizations already have years, sometimes decades, of learning content spread across LMSs, SharePoint sites, SCORM packages, videos, PDFs, PowerPoints, wikis, and knowledge bases, and they have almost no visibility into what’s actually inside all of it.
We’ve seen organizations with thousands of SCORM packages spread across multiple systems where nobody can answer fairly basic questions:
- Which courses contain outdated compliance information?
- Which content overlaps?
- Which teams already created material on this topic years ago?
- Which assets are still actively used versus effectively abandoned?
That’s not really a content generation problem. It’s a visibility problem.
That’s the problem we started focusing on with MetaLark. Not “How do we generate more content?” but rather, “How do we deeply understand the content organizations already have?”
Because once you can actually inspect learning content at scale, entirely different possibilities start to emerge. You can identify duplication, find outdated information, discover buried expertise, map skills, analyze consistency, surface hidden gaps, and reuse content intelligently instead of recreating it endlessly.
That’s a very different use of AI, and frankly, it’s one that feels a lot more valuable long term.
AI Is Becoming an Accelerator, Not a Replacement
One of the biggest misconceptions floating around right now is the idea that AI replaces instructional designers, learning leaders, developers, or content teams. I don’t think that’s what’s happening.
What is happening is that the tedious parts of many jobs are starting to disappear. The blank-page problem disappears. The first-draft problem disappears. The “summarize this transcript into a white paper” problem disappears. That changes workflows dramatically.
But expertise still matters, maybe more than ever. During the discussion, we talked about something we see constantly: people assume AI magically makes someone good at instructional design, software development, strategy, or analysis. It doesn’t. A bad process with AI is still a bad process.
Or as David put it during the conversation, “The fool with the tool is still a fool.”
That line stuck because it’s true. The people getting the best results from AI right now are the people who already understand their craft deeply. They know their audience. They know their systems. They know the business problem they’re trying to solve. AI just helps them move faster.
We May Be Moving Beyond “Courses” Altogether
This is where things start getting really interesting. For years, learning technology has largely revolved around packaging content into courses and delivering it through platforms. That model made sense. But AI agents, personalization, and contextual systems may start shifting learning toward something much more dynamic.
Instead of someone launching a course manually, future systems may recognize context and surface learning automatically in the flow of work. Not generic learning, but personalized guidance, specific help, and context-aware support. That has major implications for LMS platforms, authoring tools, and even mobile learning itself.
During the discussion, I made a somewhat provocative statement: mobile learning and microlearning may eventually “go away.” Not because learning on phones disappears, but because the idea of going into a separate app to consume isolated training content may become less important over time.
When agents understand who you are, what you’re doing, what systems you’re using, and what problems you’re trying to solve, learning becomes embedded directly into the experience itself.
We’re still very early in all of this, but it’s not hard to see the direction things are moving.
The LMS Isn’t Dead. But It Is Changing.
People have been predicting the death of the LMS for at least 15 years. That prediction still feels overstated. Organizations still need governance, tracking, compliance, reporting, and systems of record. But the role of the LMS is evolving.
Historically, LMS platforms primarily managed packaged content. Going forward, they may become much more focused on orchestration, skills visibility, activity tracking, personalization, and intelligent content management.
The same may be true for traditional authoring tools. Today’s authoring tools still largely assume the end product is a packaged course. But what happens when AI systems can dynamically assemble personalized learning experiences from underlying content instead?
That changes the role of the authoring layer significantly. Not overnight, but directionally, it’s hard not to see that shift happening.
We’re Entering the “AI Slop” Era
One of the concerns we discussed was the rise of what we jokingly called “AI slop.” Social media already has it. Learning content is next.
Just because AI allows someone to generate massive amounts of content quickly does not mean that content is useful, accurate, well-designed, or needed. In fact, there’s a real risk that organizations create even more noise than they already have: more content, more duplication, and more confusion.
More content does not automatically create more learning value. In some organizations, it may actually make the problem worse by increasing inconsistency, outdated information, learner fatigue, and duplicate material spread across multiple systems.
That’s one of the strange ironies of this moment. AI can help organizations create content faster than ever, while simultaneously making it harder to manage, govern, and understand the overall learning ecosystem if they’re not careful.
Which ironically brings us back to the original problem. Organizations don’t just need AI-generated content. They need visibility. They need understanding. They need ways to inspect and manage the content ecosystem they already have.
That’s ultimately why we believe this next phase of AI in learning is going to be less about generating more and more content, and more about making sense of the enormous amount of content that already exists.
And honestly, that may end up being the more important transformation.
About the Author
Robert Gadd is President & Chief Strategist at MetaLark and Co-Founder of OnPoint Digital. For more than two decades, he has worked at the intersection of learning technology, mobile learning, and enterprise content systems. His current focus is helping organizations finally understand what’s actually inside the massive libraries of learning content they’ve accumulated over the years.