Most AI governance discussions start too late. They start with controls, committees, policies, or compliance. This
model starts one step earlier: what level of reflection can an organization actually sustain when it judges its own
AI practice?
The core operation is simple: M(p) asks for the conditions or presuppositions of p.
Maturity is the highest level at which an organization can stably operate that question without collapsing into
theatre, paralysis, or self-congratulation.
Maturity is not how many governance artifacts an organization can display. It is the highest level at which it can
still examine the conditions of its own judgment.
01 · Why This Model
Governance is not only about controls. It is about the depth of reflection behind those controls.
This model extends the Issue Space diagnostic framework upward. Issue Space gives the
diagnostic tooling for asking whether we are even naming the problem correctly. The metacognitive maturity model
asks a broader question: what happens when institutions start reflecting on the conditions of their own diagnosis,
governance, and values?
The ladder matters because each level carries a characteristic failure pattern. At lower levels, organizations miss
systemic conditions and treat symptoms as causes. At higher levels, they risk substituting endless reflexivity for
action. The point is not to fetishize the top. It is to understand where a system can genuinely operate, and where
it is only performing sophistication.
02 · Interactive Model
Click through the levels, then stress-test the sliders against the organization you actually have.
The content and interaction below preserve the original model: five levels, expandable detail views, and a simple
assessment bar for estimating where an organization can operate with stability.
Diagnostic framework
AI governance metacognitive maturity model
A meta-recursive framework mapping organizational AI governance capability across five levels of cognitive
abstraction, from operational problem-solving to self-referential awareness.
Based on the Issue Space methodology · Shengxing Yang · with Claude Opus
Core operation: M(p) = "What are the conditions/presuppositions of p?" Maturity = the highest level at which an organization can stably operate this function.
Organizational maturity assessment
Adjust each dimension to estimate your organization's metacognitive level.
Claude Opus materially shaped the articulation of this model, so I am crediting it explicitly here.
The conceptual framing of the metacognitive ladder, the recursive logic behind the maturity levels, and the first
articulation of this model were developed together with Claude Opus. I adapted that collaborative draft into this
English essay page and interactive presentation for the site.
The broader Issue Space methodology, the book context, and the final public framing remain grounded in my own
research and writing. But this specific model benefited enough from the collaboration that omitting the credit
would be inaccurate.
Closing
What matters is not claiming a higher level. It is knowing where reflexivity still remains operational.
A model like this is only useful if it sharpens the next conversation. It should help teams distinguish between
genuine reflective capacity and decorative governance theatre, and between necessary meta-work and avoidance
disguised as sophistication.