AI does not require
leadership change.

The phrase is what you might call pure mumbo jumbo, gibberish and BS. It would be no more intellectually profound than claiming that changing the size of the paper used by a company from A4 to A3 requires a radical change in leadership.

In reality, all work and all its outputs are governed by boundary conditions, and AI does not set a single one of them. The conditions are simple:

These conditions do not change. Human needs and desires change: some disappear, some emerge and some remain. A promise must still be kept. A deadline must still be met. No model, platform or “transformation” replaces these conditions or makes them optional.

That is precisely why the idea of AI imposing “demands for change” is nothing more than a self-perpetuating circular argument:

Reliability is not a law imposed by AI. If reliability matters, the tool should be chosen accordingly. If a tool reduces reliability, there are three options:

These options arise if and when an organisation chooses a tool that brings uncertainty, errors, leaks or weak traceability. Demands for governance and change are consequences of the decision, not laws of nature, compulsions or regulations. The use of AI is not mandatory. It makes sense only if it improves the organisation’s ability to keep its promises to customers by enough to justify paying its ancillary costs. Otherwise, it is a burden.

It is not AI, then, that demands change. What calls for change are your customers, the wellbeing of your people, continuous learning and your ability to adapt. And even these boil down to the same thing: understanding what people are willing to pay for, how to deliver on the promise, and how to deliver on it on time, when agreed.

The claim that “AI requires changes to leadership and processes” is therefore thoroughly misleading because it turns the tool itself into a governing imperative. The proper order is this: customer need and the promise set the terms; technology is adopted only if it helps to fulfil them. If an organisation does not already know how to keep its promises, AI will not create that capability. If it does, AI is one tool among others — not a threshold that requires its entire approach to leadership to be redefined.

If reliability is central and the output of a generative model is systematically uncertain, its use is not inherently sensible. It makes sense only where the uncertainty can be minimised, even eliminated, where the cost is lower than the benefit obtained, and where the promise to the customer can still be kept. Otherwise, the tool should be rejected. This is ordinary tool selection, not an ontological renewal of management.

These factors do not in themselves justify a demand for change; they justify considering whether the tool should be adopted. The decision sets the consequences in motion, but the decision alone does not determine their content. That content emerges from the interaction of the use case, the criticality of the task, the characteristics of the AI and regulation. In other words: the decision activates the requirements. The use case determines what they are.

Moreover, a decision to use an unreliable or risky tool creates an entirely new family of requirements for the organisation: the decision to use a risky tool requires that the tool be governed precisely and systematically.

And that requirement for governance is a consequence of the decision; it is not a law of nature imposed by technology. The riskier tools there are, the more demand and pressure there is for governance. And the more governance, inspection and supervision there is, the more bureaucracy, time and people must be enlisted to ensure that the machine, the technology and the stages of work function as intended.

The equation is ruinous: the human being changes from a doer into an inspector, and the ugly law of nature is that the more there is to inspect, the more errors slip through inspection. This produces yet more costs, known as the “Cost of Poor Quality”, and the equation is not rescued by renaming “Final Inspection” as “Human-in-the-Loop”. It is the same old pattern: checking and bureaucracy are used to compensate for an unreliable process — and when attention is fixed on the tool, the focus shifts to models and structures for governing the tool rather than to how value is created.

If no one knows how to drive the car, there is no point in building a governance system for fuel distribution. The result is merely chaos in which a sense of control comes from knowing that at least the distribution of fuel is now under control, under the banner of “Human-in-the-Loop”, whose current form is closer to “Human-in-the-Poop”. And this is just sad, although fully avoidable.

AI itself does not require an organisation to make a single change. An organisation may decide to use AI in a way that may create a need to change processes, responsibilities or controls. But no general law follows from this: the need for change must be demonstrated, not assumed. There is no point in fixing what is not broken.

The real problem with the claim that “AI requires a change in leadership” may therefore not even be its literal falsity. The problem is what the phrase leaves hidden:

In short: AI itself does not require a change in leadership. If the foundations of leadership and value creation are sound, AI will find its place as a tool naturally and smoothly, without unnecessary drama, without vast new layers of governance and bureaucracy, without fearmongering manufactured by marketers, and without breathless, superficial AI jargon — or it will not find a place because it offers no benefit. That, too, is a perfectly valid conclusion. AI is ultimately just a tool, nothing more.

The phrase “AI requires change” is particularly misleading because it locks in the tool and turns organisational change into a quasi-compulsory default. No change derives its legitimacy from AI itself. Legitimacy can arise only if the change improves the organisation’s ability to create value and keep its promises, to move with the times, to see, to understand, and thereby to make change an inherent property rather than a separate, stand-alone capability that is “used” whenever the need arises.

The capacity to change is at the core of an organisation. It should not have to be demanded separately: it must be allowed to unfold organically and fluidly within a commonly understood operating model, moving with the times towards a shared direction. The capacity to change cannot be commanded; it simply requires the agreed, shared structures.

Move with the times. Move with Your promises. These are the only things You need.