.jpg)
There’s no doubt that generative AI will transform the way we learn how to do things—just like radio, television, and the internet did before it. This shift will provide new access to skills for billions of people. AI-assisted writing is already prevalent around the world*.
But other aspects of the AI shift are not so appealing. There’s growing evidence that GenAI encourages us to make less effort—to think less** . That’s problematic, because thinking about how we think is integrally linked to how we learn***. And making it easier for individuals to create content also leads to... a lot more content. You’ve likely noticed your social media feeds are now flooded with low-quality, AI-generated videos and images. This material is optimized for engagement but not for quality. It’s junk food content.
So how will GenAI really change the landscape for corporate learning? Will it produce more effective learning through personalized tuition, or could the effect be more malign? This article suggests that, through a plethora of AI-enabled content creation tools, GenAI might in fact worsen what we might call the “learning junk” problem.
With over twenty years of experience working across learning vendors and leading organizations in European finance, I’ve contributed to shaping approaches that emphasize effective training methods and the integration of learning with business processes.
And let me tell you: we are vulnerable to the learning junk problem—not just in terms of content creation, but throughout the whole learning ecosystem.
That’s because, all too often in my experience, L&D teams don’t pursue financial or productivity outputs that impact the business but get addicted to an “engagement high.” Clicks, likes, positive feedback. The rush is like the rush you get from a takeaway meal. It’s necessary to eat junk sometimes, but it’s not healthy as a habit. In the same way, engagement rushes have their place, but they won’t add sustainable value.
Here are some examples of what junk learning addiction looks like:
What these failures have in common is the lack of integration: between learning and business processes, between HR and learning systems, and between data sets that could reveal skill impact.
Ajay Jacob, a senior learning technology architect who has worked at Arcadis, Booking.com, and TomTom, puts it clearly:
The marketplace is crowded with glamorous solutions. I’ve seen it many times—organizations buy the latest shiny new system or content, while something fundamental, like integrated compliance training data, is missing. The junk piles up because we move from one flashy implementation to the next.
Junk learning happens when talent development teams act boldly but not wisely. They assume investment in shiny tools will lead to engagement, which will lead to skills. But each link in that chain requires deliberate planning—how content connects to processes, how systems talk to each other, and how skills are tracked.
There’s no shortcut to integrated talent technology and meaningful skills data. Like building a healthy diet or fitness plan, it’s a step-by-step process. It requires tough conversations, honest evaluation, and smart design.
Will AI Fix This?
Maybe. If used well.
But unless we fix the fundamental problem of disconnected systems, GenAI risks making it worse. New AI-powered learning platforms are everywhere. They promise faster content, personalized experiences, skills engines, and AI coaches. These technologies hold potential.
But if learning is still siloed, these tools simply make it cheaper and faster to flood the system with more junk—more binge content, more candy store skills, more sprawl, more vanity metrics.
The issue begins with how we measure success. Breaking the addiction to junk learning starts with measuring the cost of developing skills, not just content consumption.
A senior learning expert at a global bank explains:
You want to measure cost per skill, but junk tech gives you cost per content. Talent acquisition can tell you the cost of hiring for a skill. We don’t get that kind of data from our vendors. The real problem is integration. If you can’t integrate the systems, at least integrate the data.
He’s right. To truly measure skill development costs, organizations need integrated talent technology—systems that connect learning platforms with talent management and career pathing data. Even if a seamless experience isn’t possible yet, data integration is a must.
Komensky customers do things differently.
We work closely with existing tools in your learning ecosystem. Whether that’s GoodHabitz, Pluralsight, 5 Miles, LinkedIn Learning, Data Booster, Archipel, Bloomville, or Drillster. We collect both metadata and user data, offering a full data API so you can report across platforms.
Most importantly, we focus on integrated talent experiences: where skill data connects people to experts, curated content, and real growth opportunities: within the flow of work.
Komensky’s step-by-step approach was exactly what we needed for a realistic implementation. They support us in managing content, defining our skills framework, and actively engaging our employees.
The road to integrated talent technology isn’t flashy. It’s not fast. But it’s sustainable, effective, and real.
It’s why Komensky has a 0% customer churn rate.
We don’t offer junk food learning. We help you build something healthier. Something that lasts.
To support your journey, we’ve published a new guide that walks you through our process in six steps:
Nothing worthwhile is easy.
In the guide, we share common challenges, how to solve them, and real-world examples from our customers.
Download the Essential Guide to Skills Strategy for Organizations
Fedor
*AI-assisted writing is already prevalent around the world
**There’s growing evidence that GenAI encourages us to make less effort—to think less
***That’s problematic, because thinking about how we think is integrally linked to how we learn.
