Applying governance logic and organisational design principles to the problem of AI alignment — where control must be precise, constraints must create freedom, and structure must scale.
AI alignment is not purely a technical challenge. It is a structural design problem — the same class of problem that emerges whenever autonomous agents must operate within boundaries they did not choose.
Organisations have solved versions of this for decades: how to grant autonomy while maintaining control, how to create rules that enable rather than restrict, how to build systems that remain coherent under uncertainty. These solutions are transferable.
My work translates the logic of organisational governance — constraint hierarchies, decision architectures, observational feedback systems — into structural frameworks for AI alignment. The result is a set of tools that complement technical approaches with the kind of systems-level thinking that high-stakes environments demand.
Board-level documents exploring the structural foundations of AI-human alignment.
A governance-first approach to alignment, culminating in a 24-point periodic system for mapping AI-human interaction constraints.
Read paper →How structural constraints in human systems create — rather than limit — productive freedom. Implications for AI control architectures.
Read paper →Examining how feedback loops, self-observation, and restraint mechanisms operate across biological and artificial systems.
Read paper →Structural tools for reasoning about alignment, control, and decision-making under uncertainty.
A structural decision framework that sequences inquiry before action — ensuring decisions emerge from understanding rather than reflex.
Maps how layered rules and boundaries produce emergent freedom at higher levels of organisational and AI system design.
A framework for calibrating the degree of human oversight required at each stage of AI-human interaction, from full autonomy to full control.
How self-reflective feedback loops create natural governance in both human organisations and AI systems.
Drawing on experience across McKinsey, Rio Tinto, and the Fred Hollows Foundation — designing systems where governance, autonomy, and accountability must coexist.
Full background →