Five developments from the credit union AI landscape, filtered for executives who do not have time to read forty press releases. For how to act on any of them, start with the Three-Signals Test and the vendor landscape pillar.

1. Jack Henry expands into AI security

Jack Henry announced an expanded collaboration with Google Cloud to build a proprietary AI security platform for the roughly 7,400 community banks and credit unions on its systems, using Google’s agentic defense products. The same release notes Jack Henry is deploying Gemini-based agents internally, with early adopters self-reporting time savings of up to 70 percent on routine administrative tasks.

Our take: When your core vendor automates its own operations, the capabilities eventually reach your contract, so ask your Jack Henry representative what is committed versus roadmap. The security-first framing is the right one. Adversarial AI raises the stakes for every institution that outsources its core.

2. NCUA points credit unions to revised 314(b) fraud information sharing

The NCUA highlighted FinCEN’s revised 314(b) guidance, which clarifies that participating institutions can share information in real time, including video surveillance footage, IP addresses, and fraud indicators, even for attempted transactions. Participation remains voluntary.

AI-generated fraud moves across institutions faster than any one shop can detect it, which makes 314(b) registration one of the cheapest defensive upgrades available. Worth a fresh look from your BSA officer and fraud lead this quarter.

3. A $5.7B credit union puts numbers on AI fraud operations

NASA Federal Credit Union (238,000 members) earned Celent’s Model Risk Manager of the Year award for unifying fraud and AML on DataVisor’s platform: a 42 percent drop in AML false positives, 41 percent less manual review time, and a 90 percent reduction in SAR filing time through AI-assisted narratives.

The metrics are self-reported. Still, the workflow shape (one intelligence layer across fraud, AML, and behavioral signals) is the pattern worth studying, and the timeline matters: this took three years of incremental expansion.

4. Agentic payments get a standards body

The Linux Foundation launched the x402 Foundation, an open governance body for a protocol that lets AI agents send and receive payments over HTTP. Its 40 members include Visa, Mastercard, American Express, Fiserv, Stripe, AWS, and Google.

Credit unions will not build to this standard directly, but your card processors and digital banking vendors will inherit it. Add “where do you stand on agentic payments” to your next vendor review agenda.

5. Cautionary signal: AI-generated identity fraud is now the dominant threat

AU10TIX’s Q1 2026 report found a 3.89 percent confirmed fraud rate across identity verification requests, roughly 1 in 26, with synthetic identity patterns in 47.5 percent of confirmed cases and banking at 2.11 percent.

The weak point is account opening. Legacy document checks were never designed for AI-generated credentials, and if your onboarding verification stack has not been re-reviewed since 2024, that is the audit to run before adding any member-facing AI.

Get this roundup in your inbox every Thursday. Subscribe to the AiForCU newsletter.