Age can help you investigate who is missing opportunities to learn AI. It is a weak basis for deciding which course an individual needs. Choose that support from the person's work, relevant experience and current attempt, with room for them to explain what would help.
For a learning lead planning company-wide AI training, these are two separate decisions. An age breakdown may reveal a group that received fewer invitations or had less chance to practise. It does not tell you whether Sam needs help preparing an input, checking an answer or finding an approved tool.
A group pattern is a reason to investigate
The London School of Economics' 2025 report on the generational AI gap surveyed professional employees and executives across 30 countries. Employees in younger cohorts reported more training, and trained respondents reported more AI use. The report's analysis adjusted for several individual and workplace characteristics.
Those findings describe associations in a one-time survey. The authors explicitly caution that they cannot establish causation: people who already use AI might seek training. The survey does not demonstrate that an age-specific course will work, or that an older colleague has a particular skill gap.
Treat a pattern like this as a question about opportunity. Who was offered relevant training? Could they attend? Did they have access and time to use what they learned? A low-use group might need better arrangements, different material or help with a particular task. Its average cannot choose among those explanations.
Age and experience describe different things
Someone can be experienced in railway customer service and new to an AI assistant. Another person can be comfortable with chat tools while learning the railway's ticket rules. Chronological age does not resolve either distinction.
In an April 2026 Reddit discussion, a self-described hands-on engineer wanted to learn their company's AI tools while a younger colleague worried about AI taking jobs. This is an unverified personal account, with no observed training result. Its useful detail is the difference in expressed interest within one workplace. A generational label would have been a poor substitute for asking each person.
Separate what each piece of information can tell you:
| Information | Useful for | Does not establish |
|---|---|---|
Age group | Looking for differences in learning opportunities | One person's skill, interest or preferred support |
Relevant work and tool experience | Choosing a task and discussing what is familiar | Whether the person can complete and check this task |
A current attempt and the person's explanation | Identifying specific support to try next | A permanent ability level or a need shared by their age group |
You do not need to ban age from every programme review. You need to stop it doing the work of a conversation or a task observation. If you are building the wider curriculum, use role-specific learning paths to connect practice to responsibilities.
Give the same age cohort different support when needed
Suppose Sam Taylor and Jamie Lee are customer-service advisers at a regional railway. Both fall into the company's proposed 50-and-over training group. Their practice task is to draft a passenger reply about changing a ticket, using an approved sample request and the supplied ticket-change policy.
Sam knows the policy but is new to the approved assistant. He wants help choosing the permitted information to put into it. Jamie has used chat tools before, but their draft misses a restriction that applies to the requested change. Both need a useful passenger reply, but their next practice should address different problems.
Claire should also ask what Sam and Jamie find difficult and what support they want. A single attempt can reveal a useful next step without settling everything about the learner. If assessment is the unresolved part, measure AI proficiency through relevant work, with access and permitted support made explicit.
Review opportunity separately from course placement
An Urban Institute review of AI and older workers, published in January 2026, combines literature with interviews and focus groups involving training providers, researchers and industry experts. It recommends training tied to relevant work, information literacy and opportunities to practise during working time. This is qualitative guidance, rather than a comparison proving that one curriculum works best for an age group.
Those recommendations give you concrete conditions to examine. If you already use age-group information for a programme review, look at the route from invitation to practice:
- Invitation. Were eligible colleagues offered a course relevant to their work, and did they know about it?
- Attendance. Could they join within their shifts and workload, with the support they needed?
- Practice. Did they have approved access, suitable material and time to try the task afterward?
Ask people who did not participate what prevented them, before choosing a remedy. An invitation, a completed session and an opportunity to practise are different observations. Compare the same eligible population and time period when examining participation; an employer saying training is available and a worker saying they have not received it can both be accurate.
If approved access is the missing condition, use the AI access guide to address it. Reassigning that person to a beginner class would leave the access problem in place.
Reconsider one proposed age-based course
Start with one cohort you are about to place into training. Ask what evidence supports its grouping and whether age is standing in for something you could establish more directly.
- Name the work the course should help people complete and check.
- Discuss relevant experience and support preferences, then observe a suitable practice attempt.
- Group learners around the support they currently need, allowing different arrangements within the same age band.
- Review who could take part and practise, then change any arrangement that kept eligible colleagues out.
Keep the opportunity review and the placement decision visible as separate records. For Sam and Jamie, the first asks whether each had a fair chance to learn. The second explains why one practises input preparation and the other practises policy checking. Neither needs a claim about how their generation learns.
Questions about age and AI training
Should companies split AI courses by age?
Companies should not automatically assign an AI course from a learner's age. Use the task, relevant experience, current support need and learner's preferences to choose an arrangement. A voluntary group may be useful if participants want it and it addresses an identified need, but the age label does not establish that need. Review one proposed cohort against the placement steps before enrolling everyone in the same course.
What does lower AI use in an older age group tell us?
Lower reported AI use identifies a pattern to investigate, rather than a skill gap in every older employee. Check comparable invitations, attendance, approved access and practice opportunities, then ask nonparticipants what prevented them. Survey averages alone do not establish whether training caused use or which course an individual should take.
Should younger employees coach older colleagues?
Choose an AI coach who can demonstrate the relevant task, explain its checks and support the learner. Age does not establish those capabilities. A younger colleague may be a suitable coach, as may an older colleague; verify the work they can teach and the help they can provide instead of assuming technical skill or domain judgment from age.
What should we change first in an age-based training plan?
Take one proposed course and replace age-only placement with a discussion and a relevant practice attempt. Record the support each learner needs and check that they can attend and practise. Preserve useful shared teaching where needs overlap, while arranging different help where they do not. The role-specific learning guide explains how to build that practice around real responsibilities.
