← All Terms

Governance Theatre

Policies, committees and frameworks that produce the feeling of control for the people approving AI, without evidence that any control operates on the AI systems themselves.

Governance & Risk

Governance theatre is governance built to be shown rather than run. An AI policy is approved, an ethics committee is chartered, a set of principles is published, and the board minutes record that AI governance is in place. None of it produces a record of the AI systems being checked, stopped or changed.

It usually starts with fear. A board that feels exposed asks for something it can approve by the next meeting, and a document is the fastest thing to produce. The common forms are recognisable: principles that tell nobody what to do differently, a committee whose approval rate is close to 100%, a risk register in which every AI risk is amber, and a human reviewer who signs off hundreds of decisions a day and rejects almost none.

Three questions expose it quickly: show the record that the control ran last month, name the person who can stop the system and when they last did, and show a finding that changed something. Theatre survives a board meeting, which asks whether a thing exists, and fails an audit, which asks whether it worked.