Most employees do not wake up wanting to learn something at work.

They want to solve a problem, make a decision, prepare for a difficult conversation, use a new system, or avoid making the same mistake twice. Learning becomes relevant because the work requires it.

This has always been one of the most important principles of adult learning. Adults engage when they understand why something matters, can connect it to what they already know, and have an opportunity to use it.

AI has not made those principles obsolete. In many ways, it has made them more visible.

An employee can now ask for an explanation in plain language, generate an example from their own context, practice a conversation, compare approaches, or receive immediate feedback. Learning no longer has to wait for a course, a facilitator, or the next available workshop.

But access to an answer is not the same as learning.

That is the tension I find most interesting. AI can bring learning much closer to the moment of need, while also making it easier to avoid the thinking and practice that create real capability.

What Has Not Changed About Adult Learning

I would keep several principles at the center of any workplace learning strategy.

Adults need a reason to learn

The reason should be more meaningful than “this training is required.” People pay attention when the learning helps them handle a real task, reduce uncertainty, make progress, or protect something they care about.

Before building a program, I would ask: What will this person be able to do differently, and when will they need to do it?

Experience is part of the learning material

Adults arrive with patterns, judgment, assumptions, and prior successes. Good learning helps them connect new information to that experience. It may confirm what they know, add nuance, or challenge an approach that no longer works.

This is why a generic explanation often feels thin. The learner needs to test the idea against a situation they recognize.

Practice matters more than exposure

Watching a video about feedback does not prepare a manager to respond when an employee becomes defensive. Reading a guide about AI does not build the judgment to verify a plausible but incorrect answer.

Capability develops through attempts, feedback, reflection, and another attempt. Research reviews of workplace training consistently point toward learning that is connected to practice and supported through reinforcement, not information delivery alone. A review of workforce training found that effective approaches combined in-practice learning with support such as supervision, reflection, and additional learning resources.

The environment determines whether learning survives

An employee may understand a new behavior and still return to the old one if the manager discourages it, the process makes it difficult, or the team rewards speed over experimentation.

I do not think L&D can compensate for a work environment that makes the desired behavior unrealistic. Learning design and work design have to meet.

What AI Changes

AI changes the distance between a question and a useful learning interaction.

Previously, an employee might search a knowledge base, enroll in a course, ask a colleague, or wait for a manager. Now the employee can begin a dialogue immediately.

Used well, AI can make learning:

  • More contextual. An employee can ask for an explanation based on a specific role, customer, project, or level of experience.
  • More interactive. The learner can ask follow-up questions instead of accepting a fixed sequence of content.
  • More available. Practice and support can appear during the work rather than weeks before it.
  • More adaptable. The same concept can be explained through an example, comparison, scenario, checklist, or short practice exercise.
  • More private. Someone can ask a basic question or rehearse an uncomfortable conversation without first exposing uncertainty to a group.

The OECD’s work on AI and skills makes a related point: most workers will not need advanced technical AI expertise, but they will need the digital, analytical, managerial, and human skills required to use AI well. Employees who receive training are also more likely to report positive outcomes from AI adoption.

The shift is not simply from classroom learning to AI learning. It is from learning as a scheduled event to learning as a resource that can be activated inside the work.

The Risk: AI Can Complete the Task Without Developing the Person

Sometimes that is acceptable. We use tools because we do not need to master every step ourselves.

The problem appears when the organization assumes that a completed output represents increased capability.

An employee can ask AI to draft a performance review without becoming better at evaluating evidence. A manager can generate questions for a difficult conversation without understanding why those questions matter. A new employee can produce a polished analysis while missing a basic error that an experienced colleague would notice immediately.

Recent research on generative AI and self-regulated learning reflects both sides of this relationship. AI can support confidence and learning, but learners also report risks such as over-reliance, cognitive overload, and reduced effortful thinking.

I would therefore separate three uses of AI:

  1. AI does the task. This is appropriate when efficiency is the goal and the employee does not need to retain the capability.
  2. AI supports the task. The employee remains responsible for judgment while AI provides structure, examples, or feedback.
  3. AI develops the person. The tool asks questions, creates practice, withholds the answer long enough for the learner to think, and helps the person reflect afterward.

These are different outcomes. A learning strategy should not confuse them.

How I Would Design AI-Supported Learning

I would begin with a real moment in the work, not a content library.

For example, a new manager needs to address missed expectations. AI could immediately write the message for them. A stronger learning design might instead:

  1. Ask the manager what happened and what evidence they have.
  2. Help separate observation from interpretation.
  3. Ask the manager to draft the opening in their own words.
  4. Role-play several possible employee responses.
  5. Offer feedback on clarity and tone.
  6. Prompt the manager to reflect after the real conversation.

The support is still fast and practical, but the manager is doing the work that builds judgment.

I would also give employees simple control over how they learn. One person may want a concise example. Another may need a step-by-step explanation, a chance to ask basic questions, or a realistic simulation. AI makes this variation easier to provide without creating a separate course for every preference.

The important word is control, not “learning style.” People do not need content permanently assigned to a visual, auditory, or generational category. They need appropriate options for the task in front of them.

Different Generations, or Different Conditions?

I understand why organizations want a generational framework. It offers a simple explanation for visible differences in confidence, communication, and technology use.

But I would be careful with it.

The evidence for stable generational differences at work is weak. A major review of generational research argues that lifespan, career stage, and social context provide more useful explanations than labels such as Baby Boomer, Gen X, Millennial, or Gen Z.

A 24-year-old new hire and a 54-year-old executive may approach AI differently. That difference could relate to exposure, authority, risk, professional identity, experience with similar changes, or how safe each person feels admitting what they do not know. Birth year alone does not explain it.

Age can still matter. Prior experience can help an employee recognize a poor AI answer. It can also make an established habit harder to replace. A younger employee may navigate a new interface quickly while lacking the context to judge the output. An older employee may need more time with an unfamiliar tool while bringing stronger pattern recognition to the decision.

Research on teaching digital technology to older adults emphasizes relevance, feedback, collaborative learning, informal settings, and design based on learner needs. Those are not “older-worker accommodations.” They are good adult-learning practices that become especially important when confidence or familiarity is lower.

I would design for differences without turning them into stereotypes:

  • Assess current knowledge instead of assuming it from age.
  • Let people practice privately before performing publicly.
  • Use examples that respect the learner’s experience.
  • Make peer learning reciprocal. One employee may understand the technology while another understands the business context.
  • Offer support without signaling that asking for it is a weakness.
  • Measure application, not comfort with the interface.

The Role of the Manager Changes Too

When answers are abundant, the manager becomes less important as the person who knows everything and more important as the person who helps employees decide what deserves trust, practice, and attention.

Managers can make learning part of the work by asking:

  • What did you try before asking AI?
  • Which part of the answer do you trust, and why?
  • What would change your mind?
  • Where could you test this safely?
  • What did you learn that you can use without the tool next time?

This is not about monitoring every prompt. It is about creating habits of judgment and reflection.

Managers also need permission to learn visibly. If leaders present themselves as instantly competent with every new tool, employees will hide uncertainty and use AI quietly. If leaders compare approaches, admit mistakes, and discuss where AI should not be trusted, learning becomes a shared practice.

What L&D Should Measure Now

Completion rates become even less meaningful when learning can happen through a conversation, a simulation, a workflow prompt, or a short exchange during a task.

I would measure closer to the capability:

  • Can the employee perform the task with appropriate quality?
  • Can they explain the judgment behind the output?
  • Do they recognize when AI is likely to be wrong or incomplete?
  • Can they transfer the learning to a different situation?
  • Do they need less support over time?
  • Has the work improved in a way that employees and managers can observe?

The goal is not to prove that people consumed learning. It is to understand whether the organization has become more capable.

Final Thought

AI gives adults something workplace learning has often struggled to provide: relevant support at the moment a real need appears.

That is a meaningful shift. It can make learning more personal, accessible, and connected to the work.

It can also create the appearance of capability while the person becomes increasingly dependent on the tool.

My view is that the strongest learning strategies will use AI to increase practice, feedback, curiosity, and reflection. They will not use it only to produce faster answers.

The principles of adult learning have not disappeared.

The AI era is testing whether we understood them in the first place.