Survey: Banks Expect to Hand Credit and Risk Decisions to AI Without Full Explainability
FT Longitude research with SAS and Microsoft finds banks moving toward autonomous AI decisions while governance lags adoption.
New research from FT Longitude, produced with SAS and Microsoft, finds banks moving quickly toward AI systems that do not only recommend but autonomously make credit, treasury and risk decisions, while governance practices have not kept pace with adoption. The study reports that more than 90% of retail credit portfolio decisions at Tangerine are already AI-based with minimal human-in-the-loop involvement, while higher-risk portfolios deliberately retain human oversight.
The findings highlight a widening gap between the speed of AI decisioning deployment and the maturity of explainability, auditability and oversight frameworks. Industry observers note that institutions adopting autonomous decisioning without reason codes and audit trails face regulatory exposure, and that vendors able to offer built-in explainability are positioned to pass supervisory scrutiny more easily. The research was funded by SAS and Microsoft, and covers global banks rather than a specific national market.