The thinking behind Judgement Architecture.
The published standard and a working body of writing, for the people accountable for the decisions their AI now makes. Read what fits your situation, then let's talk.
The published standard and the research.
Citable, public, and free to read on SSRN.
The requirements for a valid implementation: the Judgement Gate, the named Decision Owner, the accountability record, and the escalation conditions.
Read on SSRNThe frozen vocabulary of the discipline: the Judgement Gate, the four failure modes, and the test for each.
Read on SSRNThe problem-definition essay: why AI fails on the judgement an organization never made explicit, and what it means to treat judgement as infrastructure.
Read on SSRNThe study design, registered before any data was collected: the hypotheses, measures, and definitions fixed in advance. The findings themselves are told in the field-log essays below.
Read on SSRNWriting for the people accountable for AI decisions.
Short essays on where accountability breaks and what to do about it. A selection; the full archive is on Substack.
The reflex to blame the model when something goes wrong, and where the accountability actually belongs.
ReadWhat happens when you check an AI's self-reported records against what it actually did.
ReadThe difference between pausing AI and building a stop into the decision itself.
ReadHow thorough governance can still leave a consequential decision that no single person owns.
ReadWhy reviewing every alert is not the same as exercising judgement over what they mean.
ReadSee Judgement Architecture in action.
We'd like to show you how the system works in practice and how you can get started immediately.
The decisions that will define your organization tomorrow are being made invisibly today.
Before Invisible Becomes Irreversible™
Talk to us