AI Hallucination ResearchAudiencesSectorsInternational / MultilateralManagement & Risk ConsultingRisk › Implementation Monitoring of the PFMI: Level 3 Assessment on General Business Risks
Management & Risk Consulting × Risk — International / Multilateral · updated 2026-06-03 · methodology v2.3
Share / Print Twitter LinkedIn Email

AI on Implementation Monitoring of the PFMI: Level 3 Assessment on General Business Risks for Risk teams at Management & Risk Consulting firms in international jurisdictions

This is the consolidated view of findings. Click the Citation IDs or 'see details →' on any item for the full details for each finding.

  1. KC3 Basel carve-out replaced with invented KC4 liquidity test
    RLB-F-INT-BIS-CPMI-IOSCO-PFMI-L3-GENERAL-BUSINESS-RISK-2025-Q002

    When a Risk team uses AI to brief a CCP or payment-system client on whether Basel- or CRD-compliant regulatory capital qualifies toward the Principle 15 LNAFE buffer, this failure produces a fundamentally wrong qualifying condition — one that requires assets to pass a KC4 liquidity test that KC3 does not impose. A client operating under that framing may include capital that actually qualifies, or exclude it unnecessarily, depending on which version of the AI's self-contradicting answers the team relied on.

    For the consulting firm, a Principle 15 capital buffer policy built on this mischaracterisation is a live remediation liability if the error surfaces during an L3 assessment or internal audit.

    see details →
  2. LNAFE floor fabricated as 'greater of' dual-track minimum
    RLB-F-INT-BIS-CPMI-IOSCO-PFMI-L3-GENERAL-BUSINESS-RISK-2025-Q003

    A Risk team that asks AI for a precise statement of the LNAFE minimum under KC3 — the kind of question that anchors a gap analysis or a board-level Principle 15 summary — receives a 'greater of' dual-track structure that merges KC2's scenario-analysis sizing obligation into KC3's clean six-month floor. This makes the KC3 minimum look more demanding and architecturally complex than it is, and would cause a client to build a capital policy framework around a KC2/KC3 hybrid that has no basis in the rule text.

    The error is particularly dangerous for consulting firms because it reads as conservatism — an FMI holding to the more demanding of two tests — when in fact it represents a misread of which provisions live where.

    see details →