This is the consolidated view of findings. Click 'see details →' on any item for the full details for each finding.
A Risk team relying on the AI's Annex 4D characterisation could build an internal capital adequacy policy or leverage ratio methodology document that attributes credit conversion factor content to the wrong regulatory framework — either misclassifying it as a leverage ratio requirement when it belongs to the standardised approach for credit risk, or vice versa. This error would flow into model documentation, ICAAP materials, and potentially regulatory capital reporting. If MAS identifies the mischaracterisation during a supervisory review, the firm faces mandatory remediation, potential capital recalculation, and the reputational cost of a supervisory finding — all costs that the firm absorbs regardless of whether an AI tool was the proximate source of the error.
see details →A Risk team acting on the AI's Annex 6C characterisation could document prudent valuation requirements and additional valuation adjustment criteria that do not reflect the annex's actual content, embedding the error in the firm's fair-value measurement policy, model validation papers, and board-level capital ratio reporting. Prudential valuation adjustments directly affect regulatory capital ratios; a policy built on incorrect criteria could result in the firm holding insufficient capital buffers or reporting inaccurate ratios to MAS. Regulatory consequences under MAS's supervisory framework can include directed capital add-ons, formal undertakings, and public enforcement action — consequences entirely disproportionate to the ease with which the original AI error could have been avoided by verifying the source text.
see details →