AI Is Doing What L&D Couldn’t for 10 Years. But… There’s a Catch

Rosalie Sheldon

May 4

10 min read

In 2016, the World Economic Forum declared the arrival of the Fourth Industrial Revolution and warned that millions of workers would need to be reskilled. The corporate world responded with a flood of learning platforms, skills taxonomies, capability academies, and upskilling pledges. Billions were spent. Progress was made — but slowly, unevenly, and never quite at the pace the moment demanded.

Then generative AI arrived. Every organization began investing in it to chase the same headline: productivity gains, efficiency improvements, faster time to value. But something more interesting began happening beneath the surface. Employees aren't just using AI to do their current work faster. They're using it to stretch into entirely new territories of skill and competence. The breadth that L&D had been chasing for a decade was suddenly accessible to anyone with a prompt.

For years, talent professionals have relied on McKinsey's T-Shaped Employee model as a shorthand for the ideal worker: deep expertise in one domain (the vertical bar) paired with broad collaborative knowledge across many others (the horizontal bar). Over time, the model evolved. We heard about M-shaped employees, comb-shaped employees, K-shaped employees, etc… And in each iteration, the concept grew deeper. More verticals. More specialization. More depth.

AI is doing something different. Rather than pushing employees deeper into additional verticals, it's stretching them horizontally. It allows people to acquire and operationalize new skills without needing to develop them to the same depth that was previously required. An employee doesn't need to become a data analyst to produce meaningful analysis. They don't need to become a designer to create compelling visuals. AI provides a functional layer of competence across a wider surface area. Which is, ironically, exactly what talent teams have been chasing with their reskilling and upskilling initiatives since the World Economic Forum first sounded the alarm.

But here's the catch: accessible isn't the same as durable. And breadth without depth creates a new kind of risk. The same tool that accelerates horizontal growth may quietly be eroding the verticals. AI is both helping and hindering employee skills development, and understanding this tension is essential for any organization building a workforce strategy for the next three to five years.

AI as the Helper: The "Breadth" Multiplier

Democratizing Expertise

The most immediate impact of AI on the T-shaped employee is its ability to make people functional in domains outside their primary expertise. This stretches the horizontal bar of the T-shape in ways that weren't previously possible.

The old version of breadth was knowing "a little about a lot." You couldn't do the work yourself, but you understood enough to collaborate with the person who could. The slightly more advanced version was knowing where to find the answer. A skill any good researcher or librarian's child (more on that later) would tell you is underrated.

AI adds a third layer: not only can you find the answer quickly, you can be guided through performing the task itself. A marketing professional doesn't just understand what a SQL query does. They can now write one, with AI walking them through the logic in real time. A product manager doesn't just grasp the concept of financial modeling. They can build one, iterating with AI like a colleague until it's functional. This adds a layer of practical ability to each breadth skill that simply didn't exist before.

The implication for organizations is significant. Cross-functional collaboration becomes less about handoffs and more about genuine contribution. The horizontal bar of the T doesn't just represent awareness anymore. It represents operational capability, even if that capability is AI-augmented.

Accelerated Career Mobility

For most of modern work history, pivoting careers or pursuing a promotion into a new domain required significant investment: additional education, formal reskilling programs, months or years of upskilling before you could credibly perform in a new role. Employees often felt trapped. They were aware of where they wanted to go, but unable to bridge the gap quickly enough to make the leap.

AI is changing that dramatically. According to the University of Phoenix Career Institute's 2026 Career Optimism Index, 50% of workers say AI makes them more confident about pivoting into a new role. Workers who are knowledgeable about AI report even greater optimism about available job opportunities than workers overall. When moving into a new role, employees now have a tool that helps them find, understand, and practice the skills they need at a speed and depth never before available. They can lean into AI as a bridge until they reach proficiency on their own.

The data also reveals that this isn't something organizations are driving. The employees are doing it themselves. Half of workers report they are learning to use AI independently, pointing to strong demand for AI skill-building even without formal employer support. At the same time, 47% say their employer should be doing more to incorporate AI into their work, and 60% want more guidance in learning AI tools. The appetite is there. The organizational support often isn't.

This creates a retention risk that leaders should take seriously. Nearly half of employers (48%) worry they may be unable to retain AI-fluent talent as demand for those skills grows, and 62% say employees are developing AI skills faster than the organization can adapt. Workers whose employer has a clear plan for AI-enabled growth are significantly more likely to be satisfied in their current job at 87% compared to 72% for those without such a plan. The message is clear: if you're not helping employees stretch their T-shape with AI, someone else will.

Personalized Learning Pathways

Here's where a gap is forming between what organizations are offering and what employees actually want. Most talent teams are deploying one-size-fits-all AI learning pathways focused on the mechanics of how AI works: prompt engineering basics, tool overviews, policy guidelines. These are necessary but insufficient. What employees are hungry for are personalized learning pathways that help them understand how AI can augment their specific work, in their specific role, toward their specific career goals.

The data supports the business case for personalization. According to Deloitte's 2026 Global Human Capital Trends research, organizations with personalized learning initiatives are 42% more likely to report high employee engagement showing a direct link to performance and retention. Research published in the Current Research Journal of Social Sciences and Humanities found that AI-driven L&D initiatives have resulted in 15% to 30% improvements in training efficiency and a 23% reduction in turnover among high-potential talent from underrepresented groups.

Beyond engagement, hyper-personalized learning platforms generate data that serves broader workforce planning purposes. The intelligence gleaned from these systems helps HR leaders identify emerging skills, make evidence-backed decisions about succession planning, and understand where the organization's collective capability is growing versus atrophying. Personalization isn't just better for the learner. It's better for the business.

AI as a Hindrance: The "Depth" Dilemma

As the daughter of a librarian, I've always prided myself on a simple principle: you don't have to know every answer, you just have to know how to find it. And AI makes finding answers and guidance easier than it has ever been. But here's the problem: in the way the adult brain learns, the act of going to find the answer, of reading it, comprehending it, wrestling with it, is part of the skilling journey. When finding the answer becomes frictionless, the learning often doesn't stick.

The horizontal stretch that AI enables may come at a cost to the vertical bar of the T which is the depth of expertise that has always been the employee's anchor.

Cognitive Offloading and Deskilling

When people become too reliant on AI for routine and complex tasks alike, it can lead to an ability deficit. Core human capabilities — critical thinking, problem solving, independent reasoning — begin to atrophy from disuse.

A 2025 study published in the journal Societies by Michael Gerlich found a significant negative correlation between frequent AI usage and critical thinking scores. The research revealed a strong positive correlation between AI tool use and cognitive offloading (r = +0.72) and a strong negative correlation between cognitive offloading and critical thinking (r = −0.75). In other words, the more people trust AI, the more they offload cognitive work to it, and the less they engage in the deep, reflective thinking necessary to maintain their own analytical capabilities. Younger participants in the study exhibited higher dependence on AI tools and lower critical thinking scores compared to older participants. A finding that should concern any organization investing in early-career talent.

This isn't theoretical. It's the same dynamic that has played out with every cognitive aid in history, from calculators to GPS navigation. The difference is that AI operates across virtually every reasoning domain simultaneously, making the offloading effect broader and more pervasive than any tool that came before.

The Illusion of Knowing

There's a phenomenon in cognitive science called the "illusion of explanatory depth". This is the tendency for people to believe they understand something more deeply than they actually do. AI supercharges this illusion. Employees using AI assistance often believe they have mastered a task or concept when, in reality, they've only learned to produce the output with the tool's help.

It's exactly what your middle school math teacher told you about becoming too reliant on a calculator.

A major 2026 study from Carnegie Mellon, Oxford, MIT, and UCLA provides causal evidence for this effect through a series of randomized controlled trials. The findings are striking: while AI assistance improves performance in the short term, people perform significantly worse without it. In one trial, the AI-assisted group scored 17% lower on independent post-task tests compared to those who worked without AI. That’s roughly the difference between two letter grades. These effects emerged after only 10 to 15 minutes of AI-assisted work.

Perhaps most concerning is where the erosion shows up most severely: in evaluative skills like debugging, error-checking, and quality assessment. While AI can generate solutions, users often lose the ability to fix those solutions when they break, because they never engaged with the underlying logic in the first place. In an era where AI-generated outputs increasingly require human verification, this creates a dangerous competence gap.

The researchers frame this as a persistence problem. AI doesn't just make people less skilled. It makes them more likely to give up. Participants who had used AI were significantly more likely to skip problems entirely when the AI was removed. People don't merely become worse at tasks; they also stop trying.

The Cognitive Crunch

AI allows work to become faster, but faster often becomes more demanding. When a four-hour task takes thirty minutes, organizations rarely give that time back to the employee. Instead, they absorb the efficiency into higher volume expectations — expecting sixteen deliverables instead of four, or expanding the scope of a role to cover tasks that would have previously been handled by multiple people.

This creates what some researchers are calling the "cognitive crunch." AI doesn't reduce work so much as it intensifies it, particularly when organizations use the technology to accelerate task volume without redesigning priorities, workflows, or accountability structures. A 2026 HRD Connect analysis highlighted research showing that heavy AI use can produce a pattern of mental fog, slower decision-making, and exhaustion linked to the cognitive load of managing multiple AI tools and validating their outputs — what researchers describe as "AI brain fry."

The effect is especially pronounced in decision-intensive roles. AI shifts effort from doing the work to monitoring the work. That kind of vigilance is mentally expensive, particularly when employees remain fully accountable for quality and outcomes they didn't personally produce.

Highly productive employees have always struggled with a version of this. They're so effective that organizations want to scale their output by adding team members. But when new employees can't produce at the same level, the high performer picks up the slack because "it's just easier" and then burns out. AI creates a similar dynamic at the individual level: the employee becomes the quality-control layer for an ever-increasing volume of AI-generated output, with no additional cognitive capacity to absorb it.

What Organizations Can Do

The tension between AI as breadth multiplier and AI as depth eroder isn't something that resolves on its own. Left unmanaged, organizations will end up with employees who appear more capable on paper but are more fragile in practice: wide but shallow, fast but brittle. Managing this requires intentional strategy across several dimensions.

First, have a strategy at all. The data is unambiguous: workers whose employer has a clear plan for AI-enabled growth are significantly more satisfied and more likely to stay. Yet most organizations are still in reactive mode: deploying tools without a coherent vision for how AI fits into workforce development. A strategy doesn't need to be perfect. It needs to exist, be communicated, and be iterated upon.
Second, distinguish between AI-augmented performance and AI-dependent performance. Organizations need clarity on which roles and tasks should be permanently augmented by AI and which require independent human capability as a baseline. For safety-critical decisions, creative judgment, leadership moments, and foundational skill areas, employees need to be able to perform without the tool. Training programs should include "AI-off" assessments that measure what people can do independently.
Third, invest in personalized learning, not just AI literacy. Generic AI training teaches people how the tools work. Personalized learning pathways teach people how AI changes their specific work, role, and career trajectory. The difference in engagement, retention, and performance outcomes is substantial and well-documented.
Fourth, redesign work, not just workflows. When AI makes tasks faster, the organizational response shouldn't automatically be "do more." Leaders need to be intentional about where efficiency gains are reinvested — whether that's in higher volume, deeper quality, employee development time, or recovery capacity. Without this intentionality, AI becomes an accelerant for burnout rather than a tool for growth.

Finally, treat persistence and struggle as skills worth protecting. The research is clear that AI can erode not just competence but the willingness to work through difficulty. Organizations should create environments where productive struggle is valued, where employees have opportunities to solve problems without AI as a safety net, and where the development of independent judgment is rewarded alongside productivity.

The T-shaped employee isn't disappearing. It's being reshaped by AI in ways that create genuine opportunity: broader capability, faster career mobility, more accessible expertise. But the horizontal stretch only creates value if the vertical bar remains strong enough to anchor it. The organizations that thrive in the AI era won't be those that stretch their people the widest. They'll be the ones that stretch them wisely.

Ready To Get Started?

If you're ready to start building your AI Learning Strategy, we're ready to help.

Send a quick note to rose@ckgconsulting.net and we'll connect to figure how we can get your custom approach designed and operational in as little as 90 Days.