The Compliance Translation Gap: Why AI Regulation Describes Outcomes and Organisations Need Procedures

On 27 July 2026, six days before the European Union's flagship AI legislation was due to impose its most consequential obligations, an amending regulation entered into force and moved the deadline by sixteen months. The stated reason was not that the obligations were wrong. It was that the technical standards required to operationalise them were not finished, and the conformity-assessment infrastructure the law assumed would exist had not matured. This article treats that deferral as the clearest available evidence for a structural condition it names the Compliance Translation Gap: the interval between an obligation becoming law and an executable procedure for discharging it becoming available. Legal instruments state outcomes — that a system shall be accurate, robust, and free from discriminatory effect. Organisations cannot execute outcomes. They execute procedures, assigned to named roles, producing retained evidence. Something must convert the first into the second, and in this design that something is a body of harmonised standards drafted by private standardisation organisations and cited in the Official Journal. Three constructs are developed. Translation Latency is the elapsed time from an obligation's adoption to the availability of a cited method for meeting it; computed from published dates, it runs to roughly forty months in this regime. The Presumption Void is the state in which an obligation is known and no cited standard confers presumption of conformity, so every organisation must construct its own compliance argument and bear the risk of having constructed it wrongly. Outcome-Procedure Distance grades how far a requirement sits from anything executable, and explains why some obligations translate quickly and others do not translate at all. The article identifies a control case within the same instrument. General-purpose model obligations are demanding and were not deferred, because an interim method was supplied through codes of practice rather than harmonised standards. Across three groups of obligations under one law, the variable that tracks deferral is not burden, distance, or political salience, but whether a usable method existed on the date the obligation fell due. The gap is argued to be structural rather than administrative, arising whenever a legislature regulates by outcome and delegates method to a slower body. That cost falls unevenly: large providers absorb it and sometimes help write the standard they will be judged against, while small deployers cannot absorb it and respond by deferring, guessing, or asserting compliance they cannot evidence. Four strategies are compared on cost and defensibility, a Translation Readiness Audit is proposed, four reforms are addressed to regulators and standards bodies, and the analysis is extended to four further regulatory regimes. Paper 4 of 10 in The Answerability Series. Manuscript ID COMP-TRANS-2026-04. 41 pages, 8 figures, 24 tables, five appendices including an audit worksheet, a dated chronology with verification status, and a worked trace of one obligation through the translation chain — all released under CC BY 4.0. This article is scholarly analysis and is not legal advice.