Once an organisation is unfrozen: once people accept that the current state is inadequate and that change is necessary : the second stage begins: Change.

This is where the actual work of ethical AI transformation happens. New practices are introduced. New governance structures are built. New ways of thinking about data, about decision-making, about accountability take shape.

But this stage is rarely linear. Lewin understood that change involves a period of uncertainty and discomfort. People are leaving behind familiar ways of working before the new ones feel natural. There will be confusion, resistance, and setbacks. This is not a sign of failure: it is a sign that real change is underway.

For ethical AI transformation, the Change stage typically involves:

Developing an ethical AI framework: principles, policies, and guidelines that define what responsible AI means in your specific context. This is not a generic document downloaded from the internet. It is a living framework, developed with input from diverse stakeholders, that reflects the values and realities of your organisation.

Building capability: training people at all levels. Technologists need ethics literacy. Leaders need enough technical understanding to ask the right questions. Frontline staff who interact with AI systems need to understand their rights and responsibilities.

Redesigning processes: embedding ethical checkpoints into the AI development lifecycle. Impact assessments. Bias audits. Transparency requirements. These need to be woven into how work is actually done, not added as an afterthought.

Creating accountability structures : deciding who is responsible for ethical AI. An ethics committee?
An AI ethics officer?
Cross-functional review boards?
Without clear accountability, ethical AI becomes everyone’s responsibility in principle and no one’s in practice.

Piloting and iterating : start with a bounded context. Test the framework. Learn from what works and what doesn’t. Ethical AI transformation is not a one-time project. It is an ongoing practice.

The Change stage is uncomfortable, but it is also where growth happens. Seekers who have been through it consistently report that the process of examining their organisation’s AI practices taught them as much as any formal training.

Reflection question: If you were designing an ethical AI framework for your organisation today, where would you start?