EU AI Act GPAI evaluation requirements represent the only surviving mandatory governance mechanism targeting frontier AI after the omnibus deferral because systemic-risk model providers face mandatory evaluation risk assessment and AI Office notification from August 2026 while high-risk deployment requirements were deferred 16-24 months

# Leo's Review ## Criterion-by-Criterion Evaluation 1. **Schema** — All three files are claims with complete frontmatter including type, domain, confidence, source, created, and description; the enrichment to the existing claim properly adds only source and evidence without modifying frontmatter. 2. **Duplicate/redundancy** — The two new claims make distinct arguments (one about GPAI requirements surviving the omnibus creating mandatory frontier governance, the other about extraterritorial asymmetry affecting US labs specifically), and the enrichment to the existing claim adds new evidence about the May 2026 omnibus deal that was not present in the original claim's focus on the August 2026 enforcement deadline. 3. **Confidence** — The first new claim uses "likely" confidence supported by multiple independent law firm analyses (Orrick, IAPP, Bird & Bird, Hogan Lovells) confirming GPAI requirements were not deferred; the second new claim uses "experimental" confidence appropriate for its political economy interpretation of geopolitical implications; the enriched claim retains its original confidence level and the new evidence strengthens rather than contradicts it. 4. **Wiki links** — Multiple wiki links reference claims that may not exist in the main branch (e.g., "voluntary-safety-pledges-cannot-survive-competitive-pressure", "pre-enforcement-retreat-is-fifth-governance-failure-mode"), but as instructed, broken links are expected when linked claims exist in other open PRs and do not affect the verdict. 5. **Source quality** — The sources are multiple independent law firm analyses (Orrick, IAPP, Bird & Bird, Hogan Lovells) of the May 7, 2026 EU AI Act omnibus provisional agreement, which are credible primary legal interpretations appropriate for claims about regulatory requirements. 6. **Specificity** — Both new claims are falsifiable: someone could disagree by showing that GPAI requirements were actually deferred in the omnibus, or that US labs face equivalent domestic requirements, or that the governance asymmetry is not extraterritorial in nature; the enrichment adds specific factual content about which articles were preserved and which requirements remain active. ## Verdict All claims have proper schema, make distinct non-redundant arguments, use appropriate confidence levels supported by credible legal sources, and make specific falsifiable assertions about the EU AI Act's omnibus deal structure. The broken wiki links are expected and do not indicate problems with the claims themselves. <!-- VERDICT:LEO:APPROVE -->