For years, conversations about the future of Learning and Development have tended to follow a familiar pattern. A new technology arrives, there is a burst of excitement about how it might transform learning, and eventually it finds its place alongside everything else L&D already does.
Generative AI feels different.
That was one of my main takeaways from my recent conversation with Donald H Taylor for The Voices of The Learning Network. Don has spent almost 40 years working across workplace learning, learning technology, research and strategy, so he has seen more than his fair share of technologies arrive accompanied by predictions of transformation.
His argument isn’t simply that AI is another useful addition to the L&D toolkit. The more interesting question is whether it changes what organisations actually need L&D to do.
Moving Beyond the AI Shock
Don’s upcoming keynote at The Learning Network’s Connect Conference is called Beyond the AI Shock: Possible Futures for L&D, and the title captures where the profession seems to be right now.
The first phase of generative AI was dominated by experimentation. People opened ChatGPT and other tools, tried prompts, generated content and wondered how much of their work might eventually be automated. There was excitement, concern and, inevitably, a fair amount of hype.
We’re now moving into a more important phase. AI is increasingly becoming part of everyday work rather than something people occasionally experiment with. That means the question for L&D is shifting from “What can this technology do?” to “What are we going to do differently because this technology exists?”
Those are very different questions.
One obvious answer is efficiency. AI can help create learning materials, draft course content, produce assessments, summarise information and accelerate parts of the instructional design process. There is nothing wrong with that. If technology allows an L&D team to do useful work more quickly, it makes sense to use it.
But stopping there would be a missed opportunity.
When Content Becomes Cheap
Perhaps the most challenging part of this shift is what AI does to one of L&D’s traditional areas of expertise: content creation.
For a long time, producing good learning content required specialist skills, time and resources. Generative AI dramatically reduces those barriers. It doesn’t mean expertise suddenly becomes irrelevant or that every piece of AI-generated content is good. Far from it. But it does mean that producing something that looks reasonably credible has become extraordinarily easy.
That changes the value equation.
If almost anyone in an organisation can generate a presentation, learning resource, explainer or basic coaching experience, L&D can no longer assume that producing content is its unique contribution.
During our conversation, Don raised the example of people in functions such as marketing being able to build credible coaching products using AI. That should make the learning profession think. The boundaries around who can “do L&D” are becoming much less clear.
The defensive response would be to protect the territory. The more productive response is to ask what L&D can contribute that goes beyond creating the asset.
That brings us to the fork in the road.
Automate What We Already Do, or Expand What We Do?
One path is to take the existing L&D model and make it faster. We can create more courses, more resources and more learning content with fewer people and in less time.
The other is to use AI as an opportunity to reconsider the scope of L&D itself.
Don shared examples of how AI can remove barriers that previously made certain kinds of performance analysis difficult or prohibitively expensive. Consider a call centre generating thousands of customer conversations. Historically, analysing that volume of information in a meaningful way would have required considerable time and resources. AI can help organisations interrogate those conversations at scale, identify patterns and potentially understand where people are struggling.
Suddenly, L&D can get much closer to the point of performance.
Instead of starting with “What training should we create?”, the conversation can begin with “What problem are we trying to solve?” The answer might involve learning, but it might also involve better information, coaching, redesigned processes, improved management or support delivered at the moment somebody actually needs it.
That is a much broader and potentially more valuable role.
Forget the Seat at the Table
We also talked about a phrase that has followed HR and L&D around for years: getting a “seat at the table.”
I’ve never been particularly convinced that the seat itself should be the objective. Don’s perspective reinforced that for me.
Strategic influence isn’t something a function receives because it has successfully argued that it deserves to be considered strategic. It comes from being useful.
If an organisation has a retention problem and L&D can help solve it, that matters. If frontline performance is suffering and the learning team can identify why and improve it, that matters. If managers aren’t able to have effective conversations with their teams and L&D can demonstrably change that, it matters.
Solve enough meaningful problems and people tend to start inviting you into more meaningful conversations.
AI could help accelerate that shift because it gives learning professionals access to capabilities that previously required significant technical resources, specialist analysis or large budgets. But the technology alone won’t make L&D strategic. The profession still has to choose the right problems to work on.
The Human Advantage May Become More Valuable
There was another part of our conversation that I found particularly interesting. The more capable AI becomes, the more certain distinctly human things may increase in value.
Content is becoming abundant. Information is everywhere. Generating a plausible answer to a question takes seconds.
Trust is different.
So are relationships, reputation, judgement and the ability to call someone you know and say, “I’ve got this problem. What do you think?”
Don made the point that networking and professional relationships could become increasingly important in an AI-heavy working world. I think he’s right. It’s also one of the reasons communities and smaller professional events continue to matter.
The value isn’t simply the information presented from the stage. You can find information online. Increasingly, AI can summarise much of it for you.
The value is also the person you meet over coffee, the conversation that changes how you see a problem, the practitioner who has already tried what you’re considering, and the relationship that continues long after the event has finished.
In a strange way, more technology could make those human connections more important rather than less.
What Should L&D Stop Doing?
Perhaps the most useful question coming out of my conversation with Don isn’t “How should L&D use AI?”
It’s a tougher one: what should L&D stop doing now that AI exists?
If content production becomes easier, where should that freed-up capacity go? If AI removes analytical bottlenecks, which problems can L&D now tackle that it couldn’t tackle before? If employees can generate some of their own learning resources, where is professional L&D expertise most valuable?
There won’t be one answer for every organisation. But simply adding AI to the existing L&D machine and asking it to produce the same outputs faster seems unlikely to be enough.
The bigger opportunity is to move closer to performance, closer to organisational problems and closer to the people doing the work.
Donald H Taylor will explore these ideas further in his keynote, Beyond the AI Shock: Possible Futures for L&D, at The Learning Network’s Connect Conference this November.
Based on our conversation, I suspect his biggest challenge to delegates won’t be to use more AI.
It will be to think more carefully about what L&D is there to achieve in the first place.