Action Research offers ethical AI practice a methodology that matches the complexity of its subject. AI systems are not static objects to be analysed once. They are dynamic, social phenomena that interact with human values, organisational cultures, and community realities in ways that reveal themselves only through sustained, attentive, participatory inquiry.

The cycle of Plan — Act — Observe — Reflect is not just a research method. It is a disposition — a commitment to continuous learning, honest observation, and genuine responsiveness to what is discovered.

Its core principles:

• Research conducted with people, not on them
• Action and understanding are inseparable
• Complex situations reveal themselves gradually, through cycles
• Participation is both methodologically sound and ethically required

For organisations serious about ethical AI, Action Research provides a rigorous framework for the kind of honest self-examination that real ethical responsibility requires. Not the comfort of a completed audit. The discipline of an ongoing inquiry.