Older Workers Have the AI Upskilling Edge Tech Companies Keep Missing in 2026

Tech companies have spent the last two years racing to hire prompt engineers, AI fluency coaches, and twenty-something "AI-native" generalists, and quietly pushing older workers toward the exits in restructurings dressed up as efficiency plays. New behavioral research from SHL, covered by HR Dive, suggests they are optimizing for the wrong end of the skill stack. Andy Nelesen, head of solutions and market insights at the behavioral assessment company, told HR Dive that the workers actually outperforming peers in AI-saturated roles are not the fastest prompters. They are the ones with the deepest reserves of judgment, analytical reasoning, and decision making, the human capacities that take years of real work to build.

That insight has uncomfortable implications for the way tech is currently structuring its workforce. The cohort with the strongest "root system" of durable skills is often the same cohort getting nudged into early retirement, quiet sidelining, or layoff lists. And the cohort with the weakest roots is the one being asked to carry AI transformations on top of careers that have not yet had time to grow them.

The Tree-of-Skills Framework Reframes the AI Debate

Nelesen's mental model is simple enough to draw on a napkin, which is part of why it has started circulating in HR circles. Picture a tree. The leaves are "perishable skills," the task-specific software fluencies that feel urgent right now, such as Microsoft Office macros, a particular CRM, a particular prompt structure inside a particular model. These are the most exposed to AI substitution, and they are also the easiest skills to acquire and to lose.

The trunk is made of "semidurable skills," the kind that take a year or two to build and a few years to depreciate, including cloud computing, marketing prowess, product instincts, and data interpretation. The trunk is where most upskilling budgets get spent, because it is the layer that translates cleanly into job titles and certifications.

The roots, hidden underground in most workforce planning conversations, are the "durable skills." Judgment. Critical thinking. Decision making under uncertainty. These do not show up on a transcript or a LinkedIn skills badge, but SHL's research finds they are what actually differentiate high performers in AI-augmented work. As Nelesen put it to HR Dive, "It really isn't your ability to write a great prompt in ChatGPT5...but rather a person's AI literacy and analytical ability that differentiated those who were high performers from the rest of the pack."

In other words, the workers winning with AI are not the ones who can describe the tool the best. They are the ones who can tell when the tool is wrong.

Why Tech Companies Keep Missing This

The miss is not random, it is structural. Tech compensation, recruiting pipelines, and internal mobility systems are all calibrated for visible, recent, certifiable skills, which means they over-index on leaves. A candidate who lists "GPT-5 prompt engineering" and "Claude Sonnet workflows" reads as future-ready in an applicant tracking system. A candidate who lists "twenty years of triaging ambiguous customer escalations" reads as expensive.

That bias compounds when companies start cutting. Restructurings tend to fall hardest on roles where the value is hard to itemize, which is exactly where root-system workers live. We have seen the pattern repeatedly in coverage of General Motors job cuts and AI restructuring, the GoPro workforce cuts, and the Paramount layoffs tied to the Skydance merger. The official rationale is always efficiency. The unspoken filter is often tenure.

SHL's findings suggest those cuts may be quietly eroding the very capacity AI transformations depend on. A team that has lost its senior judgment layer can ship prompts faster, but it has fewer people who can tell whether the output is safe to send to a customer, defensible in a regulatory filing, or coherent with last quarter's strategy. That is a slow-motion risk that does not show up in the first earnings call after a layoff.

What "Durable Skills" Actually Look Like in Practice

Durable skills sound abstract until you watch someone deploy them under pressure. They look like the project manager who pushes back on a confident-sounding AI summary because something in the timeline does not add up. The compliance lead who notices that a generated contract clause references a regulation that was repealed in 2024. The senior salesperson who reads a buying committee well enough to know that the AI-suggested next step would actually kill the deal.

None of those moves require typing a single prompt. All of them require pattern recognition built from hundreds of prior cycles. SHL's research, along with a separate August study from Multiverse that identified thirteen human skills associated with strong AI adoption, points to the same conclusion. The bottleneck in AI-era productivity is not access to models. It is the human capacity to evaluate model output.

That is a capacity older workers have been quietly compounding for decades, often in roles that did not call it "AI literacy" at the time. The labor force participation data from the U.S. Bureau of Labor Statistics shows workers 55 and older now make up a larger share of the prime-aged workforce than at any point in modern history, even as tech narratives keep treating them as a legacy problem to manage.

The Generational Reframe Younger Workers Need to Hear

The temptation, if you are 28 and reading this, is to assume durable skills are something you simply have to wait for. That reading is wrong, and it is also dangerous. Roots do not grow on their own. They grow because the tree is forced to reach for water, which in career terms means deliberately taking on the kind of ambiguous, high-stakes, judgment-heavy work that AI cannot shortcut for you.

Younger workers who optimize their first five years for "AI-native" leaf skills, prompt libraries, model comparisons, automation stacks, will find themselves in a brittle position by the time they hit thirty. The leaves will keep falling off as models change. Without a trunk and roots underneath, every model update becomes a re-skilling sprint.

The healthier playbook looks like the one we have outlined in our coverage of career growth strategies and building experience as an early-career worker. Take the assignment with the messy stakeholder dynamics. Volunteer for the project where the right answer is not obvious. Sit in on the postmortems even when you are not on the hook. Those are the reps that grow roots.

It is also why the workers being most threatened by AI right now are often mid-career, not late-career. They have enough trunk to be expensive and not enough roots to be irreplaceable. The escape route is to consciously deepen judgment, not to chase a newer model release.

What Employers Should Actually Be Measuring

If SHL's framework is right, the assessment infrastructure most companies use is pointed at the wrong layer. Skills inventories track certifications and tool fluencies. Performance reviews track output volume. Hiring funnels filter on recent technology exposure. None of those measure judgment, and most of them actively penalize the workers who have the most of it.

A serious AI-era workforce strategy would invert the stack. It would assess analytical reasoning before tool fluency. It would weight decision-making track records above prompt portfolios. It would treat "managed an AI rollout that did not blow up" as a more valuable signal than "completed a six-week AI bootcamp." And it would stop treating tenure as a cost line and start treating it as the root system that holds the rest of the org upright.

That shift would also change who gets retained in the next round of cuts. We have written about North Dakota workers facing AI displacement risk, Cisco's second round of layoffs, and the broader Intel layoffs in 2024. In nearly every case, the workers cut first were the ones whose roles were hardest to defend in a one-page slide. Durable-skill workers tend to look exactly like that on paper, even when they are the load-bearing columns of the team.

Practical Moves for Workers Over 55 Right Now

If you are an older worker watching the AI hype cycle from the outside, the worst thing you can do is concede the narrative. The research is genuinely on your side. The better moves are concrete.

Start by translating your judgment into AI-era language. Instead of describing a 1998 project, describe the pattern you brought to it, "rebuilt a 14-person customer ops function after a system migration," reads as durable-skill leadership to any hiring manager paying attention. Add a short line about how you would apply that pattern with current tooling. You do not need to be the fastest prompter in the room. You need to make clear you would notice when the prompt was lying.

Pair that with one or two trunk-layer upgrades. A focused certification in a current cloud or analytics stack, a hands-on stint with a mainstream AI tool inside your existing workflow, an honest portfolio of "AI helped me do this faster, here is how I checked it." That combination, deep roots plus a fresh trunk layer, is exactly what SHL's data suggests outperforms.

And keep an eye on which employers are actually behaving consistently with this research. The ones cutting senior cohorts while loudly preaching AI fluency are revealing what they value, which is leaves. The ones investing in mixed-tenure teams, internal mobility, and judgment-heavy roles are the ones positioned to compound through the next wave.

Why this advantage will compound through the rest of the decade

The durable-skills advantage for older workers is not a one-cycle phenomenon. SHL's tree framework explains why it should compound over the next five to ten years. Every new AI release prunes more of the perishable-skills leaf canopy, which means the people who built careers on prompt syntax, specific SaaS tool muscle memory, or any single-vendor certification are losing skill value with each model upgrade. The roots, by contrast, only grow more valuable as decision-making volume increases and accountability for AI-assisted judgment moves up the org chart. Add to that the macro labor-market trend covered in our 2026 entry-level squeeze piece — younger workers are increasingly being asked to operate at senior-IC productivity levels right out of school, and the ones who succeed are the ones who deliberately build judgment-and-context skills alongside the tools. For workers over 55 specifically, this means the next decade is structurally favorable if you can be visible about the judgment you bring to AI-augmented decisions. For younger workers, it means the right move is to seek out the older operators on your team and learn what they actually do all day, because the part of their job that looks slow is the part that AI cannot replicate yet.

People Also Asked

Q: Are older workers really being displaced by AI faster than younger workers?

A: Not in the way the headlines imply. The roles most exposed to AI substitution are concentrated in mid-career, mid-skill positions where the work is structured enough for models to learn. Older workers tend to sit higher in judgment-heavy roles that AI augments rather than replaces, but they are still vulnerable to restructurings that target tenure for cost reasons.

Q: What are "durable skills" in the SHL framework?

A: Durable skills are the deepest layer of the skill tree, including judgment, critical thinking, and decision making under uncertainty. SHL's research found these capacities, not prompt fluency, separated high performers from the rest of the pack in AI-augmented work.

Q: Should younger workers stop learning AI tools and focus on judgment instead?

A: No, the answer is both. Tool fluency is the leaves of the tree and still matters in the short term. The mistake is treating leaves as the whole strategy. Younger workers should deliberately seek out ambiguous, high-stakes work that forces judgment to grow underneath the tool skills, so the roots are in place before the next model wave makes today's prompts obsolete.

Related Articles

• General Motors Job Cuts and AI Restructuring
• North Dakota Workers Facing AI Job Displacement Risk
• Cisco Second Round Layoffs 2024
• Intel Layoffs 2024
• Paramount Layoffs and the Skydance Merger
• GoPro Workforce Cuts and Restructuring
• Senior Accountant Exodus Movement
• Protecting Video Game Performers from AI Threats
• Tesla Robot Training Jobs
• Unemployment Insurance Preparedness in a Recession

Future-proof your career. Metaintro tracks how AI is reshaping the job market and delivers career intelligence to your inbox weekly. Sign up free.