Treasury Outlines Six AI-Cyber Resources for Finance

Summary:

US Treasury has launched a public-private push on AI security for finance. It will publish six resources across February 2026 to help banks and insurers use AI without breaking their risk posture. The primary source is the Treasury press release: home.treasury.gov.

Scope, partners, office:

The work was coordinated through the Artificial Intelligence Executive Oversight Group (AIEOG). That group is a partnership between the Financial and Banking Information Infrastructure Committee (FBIIC) - a council of federal and state financial regulators - and the Financial Services Sector Coordinating Council (FSSCC) - an industry group focused on critical-infrastructure security.

Cory Wilson, a Treasury cyber official, is quoted in the release. Treasury’s Office of Cybersecurity and Critical Infrastructure Protection (OCCIP) serves as the Sector Risk Management Agency for finance. See the Treasury press release: home.treasury.gov.

What’s coming:

Treasury said it will publish “six resources” during February but did not list formal titles. It did name likely workstreams, so expect guidance in these areas:

• Governance
• Data practices
• Transparency
• Fraud
• Digital identity

As of February 19, 2026, no detailed schedule or document titles were posted beyond “throughout February.” See the Treasury press release: home.treasury.gov.

Mandatory or not:

These are not new rules. Treasury frames the materials as practical tools "rather than prescriptive requirements" - in other words, voluntary guidance. That said, voluntary tools often become de facto expectations in vendor reviews and procurement checklists.

Why it matters:

Banks already operate under model-risk rules and third-party risk regimes. If Treasury’s tools map cleanly to those frameworks, they will show up quickly in procurement and audit checklists.

For context, look at the Federal Reserve / Office of the Comptroller of the Currency model-risk guidance (SR 11-7) and the interagency third-party risk guidance: federalreserve.gov.

A quote to pin it:

The resources are "designed to help institutions, particularly small and mid-sized, harness the power of AI." That line comes from Cory Wilson at Treasury. Translation: these materials are meant to be practical and implementable, not academic.

Source transparency:

Initial reporting came via a short, paywalled Bloomberg Law brief that flagged the six resources and the February timing. The full Treasury press release provides the official details: news.bloombeglaw.com and home.treasury.gov.

Action items for builders:

• Bookmark the Treasury press page and subscribe for updates: home.treasury.gov.
• Map your generative-AI use cases to SR 11-7 controls and your third-party-risk playbook; close obvious gaps now. See guidance at federalreserve.gov.
• Draft a one-pager per model covering governance, data lineage, monitoring, and access controls. Assume these will be requested during diligence.
• Inventory vendors that touch AI; make sure contracts and logging are in place to support audits.

Stay ready: treat these resources as likely to influence procurement and risk reviews even if they remain nominally voluntary.