The AI Training Gap: Why Older Workers Are Being Left Out and How to Fix It

Older workers are much less likely to receive AI training than younger colleagues, and that gap is about access, not ability. To close it, make AI training the default for every employee, tie it to real tasks in each person's job, use mixed-age learning pairs, and check managers and hiring processes for the assumption that AI belongs to the young. Experienced people bring the judgment these tools lack.

How wide is the gap?

Wider than most leaders would guess. A Cypher Learning survey found that only 11% of workers over 55 had received any AI training, compared with 30% of workers aged 18 to 44. Younger workers, men and senior managers were all more likely to use AI and get trained on it (HR Dive).

A study by the nonprofit Generation, covering more than 2,600 employees over 45 and nearly 1,500 employers in the US and Europe, found only about 15% of US workers 45 and older use AI tools at work. Ninety percent of employers said they offer some training support for AI, yet the use isn't showing up among older staff.

Then there's hiring. In the same research, 90% of US hiring managers said they were likely to consider candidates under 35 for AI-related roles, compared with just 32% for candidates aged 60 and older. That's a large gap, and it is about age, not ability.

Is this because older workers don't want to learn?

That's the easy story, and the evidence doesn't really support it.

Among US workers over 45 who don't use AI, about a quarter said they're interested in learning, another quarter were unsure, and about a third said they weren't interested. That's mixed, but it's not a wall of resistance. And research published by the American Society on Aging, using survey data from workers 50 and older, argues that low AI use mostly reflects misalignment between training, job tasks and workplace practices rather than resistance or inability (Generations).

Put simply, the training often isn't designed for them, isn't connected to what they actually do, and isn't offered with the expectation that they'll take it up. Older workers already participate in training less across the board, partly because development budgets and invitations tend to flow toward people seen as "high potential," and that label skews young.

When you don't invite people, they don't come. Then their absence gets read as lack of interest.

Why does this matter for the business?

There are three reasons, and the first is the one leaders tend to overlook.

Judgment is the scarce skill. Generative AI is fluent and fast, and it's frequently wrong. It produces confident drafts that need someone who knows the domain to check them. Experienced employees are often the people best placed to spot a plausible but mistaken answer, because they've seen the edge cases. Leaving them out of AI adoption means leaving your best quality control on the sidelines.

The workforce is aging. In many countries, workers over 50 are a growing share of the labor force. An AI strategy that quietly writes off a third of your people is not much of a strategy.

Fairness and risk. Systematically excluding older employees from development can expose organizations to age discrimination claims. It also tells everyone watching that careers here have an expiry date.

What does good AI training look like?

From what I've seen work, a few principles matter most.

Make it the default. Don't wait for people to volunteer. Enroll everyone, at every level and age, in a baseline program. Opt-in programs fill up with people who are already confident.

Start with real work. Abstract courses on how large language models work lose people quickly. Instead, start with a task the person does every week: drafting a client update, summarizing a policy, analyzing a spreadsheet, preparing for a meeting. Show how AI helps with that specific thing, then let them try it.

Give time to practice. One workshop changes very little. People need protected time over several weeks to use the tools on their own work and to make mistakes safely.

Pair people across ages. Reverse mentoring and mutual mentoring work well here. A younger colleague may be quicker with the tools. An older colleague may be quicker to see when the output is wrong or when it would land badly with a client. Both learn something.

Address the fear directly. Many experienced workers worry, reasonably, that AI training is a prelude to replacing them. Be honest about how roles will change. Show how the organization intends to use AI to support people's work, and mean it.

Use language carefully. Phrases like "digital natives" and "old school" signal who the program is really for. Talk about skills and tasks, not generations.

What should managers watch for?

Managers shape who gets the stretch assignments, the pilot projects and the invitations to the AI working group. Small patterns add up.

Ask yourself a few questions. Who did I think of first when we needed someone to try the new tool? Who have I assumed "won't be interested"? When I picture our AI champions, how old are they?

If the honest answers skew young, you've found where the gap starts. Fixing it can be as simple as asking a long-tenured colleague to co-lead the next pilot.

What do the bigger surveys show?

The pattern holds in larger samples. Pew Research Center surveyed 5,273 employed US adults in October 2024 and found that most workers, 55%, rarely or never use AI chatbots at work. Use was highest among workers aged 18 to 29 (Pew Research Center).

The task breakdown is more interesting. Among workers who had used chatbots, those aged 18 to 49 were more likely than those 50 and older to use them for most tasks. For drafting written content the figures were 50% versus 38%, and for generating new ideas 39% versus 23%. But for research there was no statistically significant difference: 58% versus 53%. When older workers find a task where AI clearly helps them, they use it about as much as anyone.

To me that is the most useful finding in the whole debate. It suggests the gap closes when the use case is obvious and relevant. That's a design problem for the people running training, and they can fix it.

What does this look like in a real team?

Picture a claims team at an insurer. The fastest adopters of a new AI summarization tool are two analysts in their twenties. The person who catches that the tool keeps misreading a particular policy clause is a claims manager with twenty-six years on the job. She spots it because she has seen that clause cause disputes before.

If she had never been trained on the tool, nobody would have caught the error until a customer complained. Once she was, she became the person who wrote the team's review checklist. Most organizations have someone like her. The question is whether they're in the room when the tools arrive.

Where should leaders begin?

Look at your own data first. Break down AI training participation and tool usage by age band, along with gender and role. You may be surprised at what you find. Then set a goal for closing the gap and track it the way you'd track any other business metric.

AI is going to touch almost every job. The organizations that do best with it will be the ones that bring everyone along, including the people with decades of experience to combine with these new tools. That experience doesn't lose its worth because the tools are new. If anything, it matters more now.

If your organization is rolling out AI and wants every generation to come along, I'd be glad to help through my keynote "Ethical Artificial Intelligence With a Human Heart" or the "Intergenerational Leadership" workshop. You can reach me through the Book Ki page on siadatan.com.