Overview
- The framework adapts NIST’s AI RMF for finance with a maturity questionnaire and a matrix of 230 control objectives categorized by adoption stage to scale from community banks to large institutions.
- Treasury developed the resources through its Artificial Intelligence Executive Oversight Group led by the FSSCC and the FBIIC, with the Cyber Risk Institute serving as a co-author.
- The AI lexicon standardizes frequently used terms to improve communication across legal, technical, regulatory and business teams, and it is offered as an optional, nonbinding reference.
- Four additional releases focused on governance and accountability, data integrity and security, fraud and digital identity, and operational resilience will be published later in February.
- Controls emphasize third‑party and vendor oversight, model integrity, data quality and provenance, defined human review roles, and shutdown mechanisms for systems that show inconsistent performance.