A machine that cannot say“I don't know”will say something else instead.

The most dangerous property of a deployed AI is unearned certainty.

“I don't know” is not a failure state. It is load-bearing structure. A system that can mark the edge of its own knowledge can hand the decision to a human at exactly the moment a human is needed. A system that cannot will make the decision anyway and record it as a finding.

Verification is what that principle looks like when it is built rather than promised: a second, adversarial reading whose job is to find the edge; a human who is summoned by the machine's own admission that it has reached one; and a record of the moment it said so.

The measure of the system is not how often it is right. It is how often it correctly stops because it knew it was past its edge.

This page used to describe E.D.A.I. in the voice of a finished product. That was the same intention–reality gap the work is meant to close, so here is the actual state.

Built: the verification model (Agent, Guardian, Arbiter); the Guardian oath and induction rite; a Guardian credential minted on Hedera mainnet (one issued); two public Hedera topics created to hold verification and compliance events.

Not yet done: no verification events have been logged to those topics. No pilot has been run to completion. Every percentage that used to appear on this site was a target, not a measurement, and it has been removed.

Open problems we have not solved: recognizing uncertainty in the first place, the confident error that never flags itself; giving human verifiers enough time, independence, and expertise to actually review rather than rubber-stamp; and keeping a verification economy from rewarding throughput over judgment.

A fuller account, which supersedes the July 2025 investment white paper, is here: State of the Work, September 2026.