Reg Revans spent decades studying how people and organisations learn. His central finding was both simple and counterintuitive: traditional education: where an expert transfers knowledge to a learner: is inadequate for dealing with complex, novel problems.
Complex problems, Revans argued, require a different kind of learning. He distinguished between:
Programmed knowledge (P): existing knowledge that can be taught, transferred, looked up. Textbooks. Frameworks. Best practices. This knowledge is essential but insufficient for genuinely novel challenges.
Questioning insight (Q): the capacity to ask fresh questions, to challenge assumptions, to see a situation differently. This cannot be taught in the conventional sense. It develops through practice: specifically, through the practice of engaging with real problems alongside others who are doing the same.
Revans expressed this as a formula:
L = P + Q.
Learning equals programmed knowledge plus questioning insight. In a world of rapid change and genuine complexity: which describes the world of AI ethics precisely: Q becomes progressively more important.
Action Learning is the methodology Revans developed to cultivate Q. It brings a small group: typically 4–8 people: together to work on real, important, unresolved problems. The group is called an Action Learning Set. Its members are called Set Members. Their role is not to solve each other’s problems. It is to ask questions that help each person develop their own understanding and find their own way forward.
The learning in Action Learning comes from three sources:
– The problem itself: engaging with real complexity produces real insight
– The questions of others: being asked the right question at the right moment can shift understanding profoundly
– Reflection on action taken: what happened, what it means, what to do next
Reflection question: Think of a genuine ethical AI challenge you are currently facing. Is it a puzzle: with a known solution that you need to find: or a problem: genuinely unresolved, with no clear answer? How does that distinction change what kind of learning you need?