Writing on formal AI governance.
The case for deterministic enforcement over probabilistic hope—in plain English, with evidence.
Why AI governance evidence must exist at decision time, not reconstructed after
Log archaeology is not evidence. When a regulator asks why an action was allowed, the record needs to have existed at the moment of decision—not assembled from scattered systems the week before the examination.
Read →The 0.1% problem: why probabilistic AI cannot govern itself in regulated workflows
An AI that follows compliance rules 99.9% of the time violates them 0.1% of the time. In a high-volume regulated workflow, that is not a rounding error. Probability is not a defence.
Read →How to prove an AI agent followed the rules: a technical architecture guide
The difference between a log that says what happened and a proof that shows the rule held before the action ran. Architecture, not monitoring.
Read →Formal specification vs. iterative debugging: what changes when you specify before AI generates
Testing tells you what happened in the cases you thought to test. Formal specification tells you what can happen across every possible input. In regulated AI, the difference is not academic.
Read →