Training Was Never About Knowledge

I had a conversation yesterday with a group of instructional designers who had been told something that left them rattled: in a world where anyone can retrieve any piece of information through ChatGPT, Claude, Copilot, or whichever model they happen to favour, there will be no reason for workplace training to exist, nobody will ever need a course again, and they should probably start thinking about which industry to move into next.

Paid subscribers will receive an additional fortnightly omnibus edition containing an additional 5-6 articles, along with early access to white papers and special deals on my upcoming L&D field guide book series. Annual membership is just £35!

What surprised me wasn’t the claim, which gets recycled every few days, but how little resilience the team had in defending what they do. My response was that the claim rests on a misreading of what L&D is for. We were never knowledge management, and if your courses exist only to hand over information that someone could find on a search engine or surface from an LLM, then yes, you should stop building them; the trouble is that you shouldn’t have been building them in the first place.

A course is the place where knowledge becomes capability. People can look up whatever they want, whenever they want, and almost wherever they want; the ability to apply that information under realistic conditions is not something they can retrieve from a prompt. Competent application is built through structured practice, reflection, and feedback, and providing those conditions, in a safe environment where people can rehearse and develop skill before carrying it into the working world, is the work that separates a course from reading the knowledgebase.

The people making these sweeping declarations are usually pushing an agenda, or stripping out the nuance for the sake of a headline. None of which makes the tools irrelevant. I suspect the era of pre-constructed branching pathways is coming to a close, as AI makes far more interesting, multi-dimensional, and personalised routes to practice possible; I also think the current generation of LLMs is too error-prone, too inclined to wander off, to be trusted at organisational scale today, and that’s fine. When technology is promising but not yet ready, the sensible response is to wait before rolling it out widely, whilst keeping your LMS and your course library exactly where they are.

So the more useful question is:

What a training experience we can offer that lets people practise, contextualise, and develop into confident, competent performers once the knowledge itself is a search away?

For all the noise about everything changing, this reponsibility has remained constant.