Our Take
Vendor-led upskilling is defaulting to tool training; the people who need it most (older workers, those without formal tech backgrounds) are being left behind by assumption, not evidence.
Why it matters
Companies betting on AI adoption are treating it as a software problem when it's a behavioral one. Exclusion by accident during upskilling now means exclusion by design in your workforce later.
Do this week
L&D leads: audit your current AI training curriculum this week and identify which modules assume prior technical knowledge or exclude older cohorts by design, so you can retrofit before rollout.
The gap in AI upskilling programs
Most corporate AI training initiatives focus on technical integration: how to use a tool, how to prompt a model, how to integrate an API. A workplace behavior expert cautioned HR Dive that this narrow focus is structurally flawed. Training that skips soft skills, change readiness, and age-inclusive design will fail adoption, regardless of how well the software works.
The warning surfaces a pattern already visible in workforce AI pilots. Teams that succeed do more than hand out access credentials. They teach employees how to think differently about their work, how to trust (or not) AI output, and how to maintain domain judgment when a system offers suggestions. These are not software problems.
Older workers and those without formal tech credentials are disproportionately at risk. Not because they cannot learn, but because training curricula assume a baseline of familiarity that doesn't exist and rarely scaffold from first principles. The result: partial adoption, skill gaps that widen with tenure, and a false narrative that older workers resist AI when the training design itself was never built for them.
Adoption failure compounds
When upskilling misses behavioral foundations, companies don't just waste training budget. They create two tiers of AI literacy within the same team. Junior staff with recent tech exposure get it running; experienced employees either opt out or produce incorrect outputs because they don't understand when to trust the tool. Both are worse than not using it.
Age and background diversity in AI readiness also becomes a retention and morale issue. If your 15-year tenure employee feels left behind by a tool everyone else "gets," they either leave or disengage from AI-adjacent work. You lose judgment and institutional knowledge exactly when you need it most to catch AI errors.
What changes in upskilling design
Effective AI training separates three layers: the tool (software), the thinking (how to work with an AI system), and the decision (when to override it). Most programs compress all three into "use the feature." They should be taught separately, with soft-skills content front-loaded.
Age-inclusive design means starting simpler, not dumbing down. Scenario-based practice before live use. Peer learning where experience matters (domain knowledge is an asset in AI contexts, not a liability). And explicit permission to say "I don't trust this output yet."
HR and L&D should also measure adoption by cohort, not company average. If older workers adopt at 60% of younger workers' rate, your training design is the variable, not their capacity. Audit the curriculum with that lens before the next rollout.