Lewin’s three-stage model gives us a coherent way to think about ethical AI transformation:not as a one-time project or a policy exercise, but as a genuine change process with distinct stages, each requiring its own approach.

Unfreeze: create the conditions for change by surfacing the insufficiency of the status quo.

Change: build the frameworks, capabilities and processes that embody ethical AI in practice.

Refreeze: embed the new practices so deeply that they become the organisation’s new normal.

And throughout: Force Field Analysis as a practical tool for understanding and navigating the forces at play.

One mind cannot see it all. Ethical AI transformation requires the wisdom of many — diverse perspectives, honest conversations, and the courage to change not just systems, but the assumptions those systems are built on.