Every board that takes AI seriously eventually arrives at the same question, and it is the right one: how do we know our governance is good enough? It is asked in earnest, and it is usually answered badly, with a count of policies written or committees seated. Those are inputs. The question is about a result.
The Center's answer is a single standard. When a regulator, a court, or an auditor asks whether the organization made an informed decision about AI risk, the program must be able to demonstrate the answer in evidence. We call it the Informed Decision Standard, and it reframes the entire exercise.
The question is not whether you have policies. It is whether you can prove you decided well.
Three conditions
The standard is satisfied when three things are simultaneously true.
First, the Governing Body had what it needed. The information required to make oversight decisions about AI actually reached the people accountable for making them, in a form they could act on, in time to matter. A risk that surfaced only after the harm does not meet the standard.
Second, the program is real rather than ceremonial. The policies on paper correspond to what happens in practice. A charter that no one can act on, a tolerance no one enforces, an escalation path no one has ever used, these are artifacts of governance theater, and a serious examiner will see through them.
Third, the chain is traceable. The line from policy to practice to assurance can be followed in evidence. One can start at a board-approved principle and trace it down to the control that implements it and the assurance that tests it, or start at an operational decision and trace it up to the authority that permitted it.
Why a standard, not a checklist
A checklist asks whether specific items exist. It can be satisfied completely while the organization remains unable to govern AI at all, because existence is not the same as function. The Informed Decision Standard asks a harder and more honest question: not whether the parts are present, but whether they add up to a decision the organization can stand behind.
This is deliberately demanding, because the audiences that will eventually apply it are demanding. A court applying the duty of oversight, a regulator examining a governance program, an auditor providing assurance, none of them will be satisfied by a binder. Each is, in its own language, asking the same thing the standard asks.
What it changes
Adopting the standard changes what a board asks for. Instead of "show me our AI policy," the question becomes "show me, in evidence, that we made an informed decision about this risk." That single shift redirects effort away from producing documents and toward building the capability the documents are supposed to represent. It is also, not incidentally, what allows an organization to move quickly. A board that can demonstrate an informed decision after the fact is a board that can make one with confidence in the moment.