·   AI has become a top emerging risk on every bank boardroom agenda

·      Predictive AI offers immediate benefits in balance sheet optimization and customer behavior modeling

·     Generative AI use remains narrow due to governance gaps and risks like hallucinations

Existing SR 11-7 model risk frameworks cover Predictive AI but need adapting for Generative AI

·      Impact-based tiering and sandbox testing are key guardrails for responsible Generative AI adoption

·      Automation is reshaping workflows, shifting value from data collection to strategic interpretation

·      Banks must avoid “model definition arbitrage” to prevent hidden risks while encouraging innovation

·      Trust-ready adoption requires balancing Predictive AI’s maturity with cautious Generative AI exploration