The Same Old Thing, Now With AI

The conversations we are having about artificial intelligence in L&D and HR are, by and large, disappointing. Not because the technology lacks potential; it is remarkable in ways that are still unfolding. The disappointment lies in what we are choosing to do with it.

We see AI-powered content creation tools that produce what we were already producing, only faster. We see agentic implementations that describe themselves as cutting-edge, where the agents are completing the same processes our teams have always followed, just without the people involved. We see organisations announcing that AI has allowed them to scale their onboarding programmes and their compliance training at volumes that were previously impossible. And in each case, when you look carefully at what has changed, the answer is: very little. The output is faster, the volume is higher, and the methodology is identical to what we were doing a decade ago.

We do not have a long or distinguished track record of delivering on their promises to the business. The research on whether our interventions translate into meaningful performance change is far from flattering, and an honest reading of that evidence suggests a significant proportion of what we have designed and delivered has had limited effect on the performance problems we were nominally trying to solve.

Given that history, the most rational response to a transformative technology would be to ask whether we could use it to do something fundamentally different. To ask what was previously impossible that is now possible. To interrogate our assumptions about what L&D is for, and whether AI might enable us to move closer to genuine support at the point of need, to more rigorous discovery of what performance requires, or to forms of intervention that adapt to context in ways that no catalogue of eLearning modules ever could.

I am not seeing that conversation in any meaningful volume. I am not seeing it in the tools being built, in the case studies being shared, or in the professional discussion taking place around us. What I am seeing is volume: more content, faster and at lower cost, usually with lower quality to boot, with agents replacing the people who used to make it. The implicit assumption beneath all of this seems to be that if we can do more of what we have always done, and do it cheaply, the impact will follow. There is no serious reason to believe that it will.

We have arrived at a moment where we ought to be questioning our purposes and methods with some urgency, and the industry response has been to automate the status quo. That is a choice we are making, and we should be clear about what it says about us, and about what we believe L&D is here to do.