Machine-readable rules are regulations, codes, or policies encoded as executable logic — with effective dates, jurisdictional scope, and citations preserved — so that compliance can be evaluated mechanically and repeatably. The rule remains law; the encoding makes it computable.
Good encodings preserve provenance: every executable clause links back to the natural-language provision it implements, so a verdict can always cite its source. Amendments become versioned data changes rather than software releases, and historical determinations stay replayable under the rule version that governed them.
The frontier of this work is public: governments digitalising construction codes, electoral statutes, and financial regulations into formats that both their own systems and vendors' systems can execute against.
This vocabulary is implemented, not aspirational — it describes how the MAIA substrate actually runs. See the platform or read the research.