Governance theater is a specific failure mode in AI portfolio management that tends to emerge in mature governance frameworks where process has expanded beyond the capacity of the organisation to act on its outputs.
The diagnostic test is straightforward: what decisions have changed because of this governance activity in the last six months? If the honest answer is "none" — if the reports are produced, the reviews occur, and the same investments continue unchanged regardless of what the review finds — then governance has become theater.
This pattern is distinct from governance that is actively developing. Early-stage governance that is not yet changing decisions may still be building the information base, establishing ownership, and creating the conditions for better decisions in the future. The theater pattern is characterised by stability: the same process, the same outputs, the same outcomes, quarter after quarter.
Governance theater is particularly common in three situations. First, when the governance function lacks the authority or political support to act on negative findings. Second, when the investment portfolio has become so large and socially distributed that stopping any element requires a level of organisational conflict the governance body is not willing to sustain. Third, when the governance process was designed to demonstrate that governance exists rather than to improve the decisions being made.
The most effective remedy is a governance audit that asks explicitly what decisions have changed, what findings are persistently unresolved, and whether a lighter governance process would produce the same useful outputs at lower cost.
For context on proportionate governance design, see The Cost of AI Governance.