DemoPricing

De Facto vs De Jure Rules in AI Governance

De jure rules are the rules as written. De facto rules are the rules actually followed, visible only in the decisions themselves. Where they differ, an institution is governed by rules nobody approved.

Where de facto rules come from

Informal exceptions that became habit, thresholds adjusted to cope with volume, cases handled one way at one branch and another way elsewhere. None is written down, so none is reviewed.

Finding them

With decision logs, Studio infers the thresholds and patterns decisions actually follow and sets them beside the written rules: where they agree, where practice is stricter or looser, and whether the difference changed outcomes.

What to do with them

A de facto rule that works can be written into the policy and sealed, so it is reviewed and enforced like any other. One that does harm can be stopped, with an escalation for the cases it covered.

Example

Scenario

A leave policy requires two weeks' notice for annual leave.

Without governance

Managers routinely approve shorter notice for some teams. The written rule says one thing and the decisions another.

With NOMOS

The comparison shows the pattern and the teams it applies to. HR writes the exception into the policy, and the sealed rules now match what is done.

Frequently asked questions

Related concepts

See it in practice

Upload a policy, confirm its rules, seal them: any system can then check an action against them.

Open Studio →