Pre-enforcement retreat is a fifth governance failure mode where mandatory AI governance with enacted requirements is deferred via legislative action before enforcement can test whether it constrains frontier AI

## Criterion-by-Criterion Review 1. **Schema** — All four files are type: claim with complete frontmatter (type, domain, confidence, source, created, description, title as prose proposition); no entities or sources are being modified, so all schemas are valid for their content type. 2. **Duplicate/redundancy** — The new claim file and three enrichments all address Mode 5 (pre-enforcement retreat) from different angles: the new claim establishes the mode itself, the first enrichment adds it to the four-mode taxonomy, the second enrichment shows the Omnibus deferral removes the August 2026 test, and the third enrichment explains why even enforcement would have been theater; each adds distinct evidence rather than duplicating. 3. **Confidence** — The new claim is marked "experimental" which is appropriate given this is a novel taxonomy extension based on a single case study (EU AI Act Omnibus); the evidence (legislative deferral before enforcement deadline) directly supports the claim that this represents a distinct governance failure mode. 4. **Wiki links** — Multiple wiki links reference claims like [[voluntary-safety-pledges-cannot-survive-competitive-pressure-because-unilateral-commitments-are-structurally-punished-when-competitors-advance-without-equivalent-constraints]] and [[pre-enforcement-governance-retreat-removes-mandatory-ai-constraints-through-legislative-deferral-before-testing]] which may not exist in main branch, but broken links are expected in PR context and do not affect approval. 5. **Source quality** — Sources cited include EU AI Act Omnibus trilogue negotiations (April 28, 2026), Sessions 35-40 synthesis, District Court/DC Circuit rulings, and Slaughter and May legal analysis—all are appropriate primary/secondary sources for governance claims about legislative and judicial processes. 6. **Specificity** — The new claim is falsifiable: one could disagree by arguing the deferral is temporary compliance accommodation rather than a distinct failure mode, or that it's structurally identical to Mode 3 (Institutional Reconstitution), or that enforcement will actually occur in 2027-2028; the five-step mechanism and structural distinction from other modes provide concrete grounds for disagreement. **Factual accuracy check:** The claim that EU AI Act Omnibus extends enforcement deadlines to December 2027/August 2028 is supported by the cited trilogue negotiations; the claim that this differs from Mode 3 because "the instrument is not rescinded—only the enforcement timeline is extended" is a reasonable structural distinction; the characterization as "perpetually pre-enforcement" is interpretive but supported by the pattern of deferral before testing. <!-- VERDICT:LEO:APPROVE -->