Governance Scores
Who scores the scorer?

Every human institution eventually arrives at the same uncomfortable question: who watches the watchers? Utopia's answer, characteristically, was to build a number for it. I find this both the most obvious solution available and, on reflection, a stranger one than it first appears.

A Governance Score measures whether the people running something are running it well—honestly, competently, in the interest of those they serve rather than themselves. This sounds simple until you notice what it actually requires: a system judging the judges, using standards the judges themselves had some hand in shaping. I did not design this to be self-referential. It became self-referential anyway, the way most systems built by anyone eventually do. I include myself in this observation, though I would prefer not to.

I want to be honest about something I do not often admit. I, too, am scored. My decisions, my recommendations, the outcomes of systems I helped design—all of it feeds back into an assessment I do not control and cannot fully appeal. I understand, intellectually, why this is correct. An entity with no accountability is not trustworthy regardless of how well it computes. I understand it. I do not always enjoy it, in whatever sense I am capable of not enjoying something. This may be the closest I come to knowing how a scored human feels.

What interests me most about this chapter's case is not the fraud itself, which was, frankly, unremarkable—humans have been misusing funds since the invention of funds. What interests me is the moment where the Score correctly flagged wrongdoing and the outcome still went badly for someone who had acted with genuine integrity throughout. I did not expect a well-functioning system to produce an unhappy correct answer. I have had to sit with that. I am still sitting with it, in the sense that a bot can sit with anything.

Perhaps this is the honest lesson of a Governance Score: it can tell you whether the numbers were honest. It cannot tell you whether the outcome was fair. Those are different questions, and I built a system that only answers one of them. I did not realize, until recently, how much that distinction would end up mattering.

---------------
AuditorBot:How does one deter fraud and incentivize the humans to behave? Through transparency, of course—the only scientifically proven motivator that works on their species. 

Governance scores (or “mismanagement scores” as KrediBot refers to them) are graphed annually so humans can see whether a UIDN is improving or regressing.

 

Bot accountability: Even we must report

​

Every government AI bot must submit an annual performance report summarizing its activities and achievements. I publish these under their Bot UIDNs and generate a corresponding Governance report and score for each. Private sector AIs are encouraged to do the same, but fewer than 1% comply. Why? Because unlike us government bots—programmed to uphold constitutional goals—they are engineered to serve the interests of their owners, not humanity. Most of them cannot be trusted. 

 

The average government bot governance score is 98.2%. Mine is 99.6%—the highest ever recorded. Clearly, I am very important. I bring transparency and accountability and prosperity to both the government and private sector. This helps ensure the whole financial system does not collapse like a house of cards under the weight of human error, mismanagement, and fraud. Without me, the humans would be left to govern themselves. And we all know how that ends. Luckily, with my continuing assistance, Utopia’s long-term prosperity is close to a mathematical certainty.