Human Development in the age of AI

Nial O'Reilly • November 10, 2026

Why the most important distinction in leadership just became even more important!

Complicated is for machines, complex is for us

There is a distinction I come back to in almost every piece of leadership work I do, and it has never mattered more than it does now. It is the difference between a problem that is complicated and one that is complex.

A complicated problem has a lot of parts, but they behave and are predictable. Given enough expertise, data and time, you can analyse cause and effect, find the answer and repeat it. Building a jet engine is complicated. So is a tax return, a supply-chain model or a regulatory submission. Hard, yes. But knowable.

A complex problem is different in kind. Its parts are people, relationships and histories, and they keep changing in response to each other and to you. Cause and effect only become clear looking backwards. You can’t solve it once and file the solution, because the moment you act, the system shifts. Raising a teenager is complex. So is changing a culture, merging two leadership teams, or deciding how your organisation should respond to AI itself.

My colleague David Rooke, writing about medical leaders, put it bluntly: ‘most leaders fail to differentiate the complicated from the complex, and so approach problems inappropriately.’ In my experience that is as true in a boardroom as it is on a hospital ward.

“Everything in our business is complex”

A Harthill colleague told me about a participant on a programme who insisted that everything in his business was complex. She asked him whether his problems were really complex, or merely complicated. He tried to wave it away as just language. She held the line.

I recognise that participant. The word “complex” has become a badge, a way of saying this is hard and important. But when we call a complicated problem complex we let ourselves off the hook of rigorous analysis. When we call a complex problem complicated we do something more dangerous. We go looking for the right answer, commission the expert, build the plan, and are then baffled when the system doesn’t do what the plan said it would.

What AI changes

Here is why this distinction has suddenly become more urgent. AI is extraordinary at complicated. Give it the parts, the data and the patterns, and it will analyse, model and draft faster than any team I have ever worked with. It is mind blowing to me and it’s a real gift. It frees people from a great deal of work that used to consume them.

But AI is built on what has already happened. It predicts from patterns in the past. That is precisely what complicated problems reward and complex problems punish. In a complex system the past is a useful guide but never a reliable map, because the system is learning and adapting too. The risk isn’t that AI gets complex problems wrong. The risk is that it gets them wrong confidently, in fluent, well-structured prose with a recommended option at the bottom. It makes a complex problem look complicated. And a leader who has built a career being rewarded for having the answer will be very tempted to take it. So the old failure to tell complicated from complex doesn’t go away with AI. It speeds up alarmingly.

What complex asks of us instead

The complexity researcher Jean Boulton makes a point I find helpful: complexity doesn’t mean sitting back and waiting to see what emerges. It means acting with intent and adapting as you go. Experimenting and adapting. Sensing and responding. Analysing and noticing. Committing to a direction while running small experiments at the edges, and being ready to drop what isn’t working. Opening up all channels of awareness and ways of attending.

She also talks about how we get “held in the polarity”: we think the decision is about which option is right, rather than working with the upsides and downsides of both. That is a very different stance from “find the answer”. It asks leaders to hold a question open long enough for the situation to show them something.

In Harthill’s language, this draws on several of what we call the Transformational Capabilities:

Inquiry-Based Experimentation: staying curious and learning by trying things, rather than relying on fixed answers
Reframing Mind: seeing the same situation through different lenses and holding apparent opposites
Complex Systems Leadership: sensing patterns and interdependencies, and asking “What am I not seeing?”

None of these is a technique you can add on. Each is expressed through, and limited by, the way a person currently makes sense of the world. That is why developing them is vertical work, and why it matters so much now.

The good news

There is something hopeful in all this. If AI takes on more of the complicated work, and does it well, leaders and teams get time and attention back. The question is what they do with it. They can spend it being faster experts. Or they can spend it on the work only they can do: making sense of the complex, together, in relationship with the people and systems involved. I have seen senior teams discover that this is not a lesser kind of leadership but a more demanding and more satisfying one. In the words of Bayo Akomolafe: ‘times are urgent; let’s slow down.’

A question to carry

Take the three biggest issues on your team’s agenda right now. For each one, ask honestly: is this complicated, or complex? Then ask: are we treating it the way it deserves to be treated?

Next in the series: AI can bet. It cannot make meaning.

< Older Post

Human Development in the age of AI

By Nial O'Reilly • November 9, 2026

Why AI makes leadership development more urgent, not less

Answering the Business School Challenge

By Nial O'Reilly • July 22, 2026

Over 10 years of vertical development on the University of Galway Executive MBA

The medical leadership gap Europe is about to fill but perhaps the wrong way!

By Nial O'Reilly • June 8, 2026

A new European survey of psychiatric trainees , published in March, lands a finding that should not surprise anyone and yet keeps surprising the system: 63% had no leadership training in their programme, every single respondent said there should be, and those who did receive some were, on average, dissatisfied with it. The diagnosis is correct. The treatment being prescribed worries but does not surprise me. Look at how the survey — and the institutions responding to the same gap — define “leadership.” It is a list of competencies to be acquired: communication, teamwork, public speaking, professionalism, networking, media interaction. The EPA’s new Leadership Academy, launched last year and a genuinely welcome initiative, describes its aims in almost identical terms: enhancing communication skills, team management, presentation and public speaking technique. Even the UEMS mandate, which I’m glad exists, asks trainees to “acquire and continuously maintain” leadership skills. Acquire and maintain. Those are horizontal verbs. They describe the addition of capabilities to a person whose way of making sense of the world is treated as fixed. But notice what the survey authors keep tripping over without the language to name it. They report that the transition from trainee to consultant — the increase in responsibility, and crucially the increase in “awareness about limitations” — changes which skills people value, and they cite this as a driver of leadership development. That is not skill acquisition. That is a shift in how a person constructs meaning, authority and their own role. In the language of adult constructive-developmental theory, and what we recognise in Harthill a s movement between Action Logics — the kind of vertical growth that no competency checklist can deliver, because the checklist assumes the meaning-maker stays the same while only the toolkit grows. This matters beyond psychiatry. The paper itself notes that the COVID pandemic was what “firmly established” leadership as central to medical identity. Yet what that period actually exposed in healthcare leaders was rarely a deficit of communication technique. It was the struggle to hold complexity, ambiguity and competing legitimate demands without collapsing into false certainty or paralysis. That is a capacity of mind, not a line on a syllabus. You can be an extraordinarily skilled communicator and still be overwhelmed by a problem that exceeds your current way of organising reality. So here is the risk. Europe has correctly diagnosed that medical leadership training is absent. The instinct now is to fill the vacuum quickly, and the fastest thing to reach for is a competency framework — definable, teachable, assessable. We will end up training the what of leadership while leaving untouched the who that determines whether any of it transfers under pressure. None of this is an argument against the EPA Academy or against teaching skills. Skills matter. The point is sequencing and depth: horizontal capability sits on a vertical foundation, and a leader’s stage of development shapes whether they can even use the skills we give them. The same conflict-resolution training lands entirely differently on an Expert-stage clinician defending technical correctness and an Achiever-stage one balancing outcomes across a team — and differently again on someone beginning to question the systems they operate inside. The opportunity for the medical leadership community is to build the vertical dimension in from the start, rather than discovering in ten years that the competency programmes plateaued. The evidence is already pointing there. The survey’s own most-favoured modality — a blend of theory and lived practice, reflected on over time — is exactly the experiential, inquiry-based structure that supports developmental growth, and exactly what most current provision lacks. Diagnose the gap as horizontal, and you fill it horizontally. Recognise that what complexity demands is bigger minds, not just fuller toolkits, and you build something that lasts. Santos et al., “Leadership skills training needs of early career doctors: a European survey,” Academia Mental Health and Well-Being, 2026.

Coaching, supervision and the thing with feathers

By Ian Mitchell • February 5, 2026

Taking inspiration from Emily Dickinson and the HopePunk concept, Ian Mitchell uncovers a series of ‘contrapuntal weavings’ for coaches and supervisors to ponder in their practice