On March 23, 2026, the U.S. Department of the Treasury announced the AI Innovation Series, run jointly by the Office of the Financial Stability Oversight Council (FSOC) and Treasury’s Artificial Intelligence Transformation Office (AITO). The format: four roundtables that put financial institutions, technology firms, regulators, and domain experts at the same table, with the stated aim of identifying high-value AI use cases and finding practical ways to scale them without compromising safety and soundness.
This is not another policy listening tour. Treasury Secretary Scott Bessent framed AI leadership as “a crucial component of economic security” and said the regulatory posture is shifting — away from constraint and toward treating the failure to adopt productivity-boosting technology as a risk in its own right.
What the AI Innovation Series Is
According to the press release, the series is built around four roundtables with four constituencies: financial institutions, technology firms, regulators, and specialized experts. The agenda has two threads. One is mapping where AI actually runs in finance — fraud detection, cybersecurity, credit underwriting, and operational risk management are already deeply AI-embedded functions. The other is finding practical routes to scale innovation, including the regulatory friction that currently slows bank AI adoption.
The official goals: make governance, supervision, and market practices evolve with AI capabilities; keep the U.S. financial sector ahead on AI adoption while protecting national security and long-term economic resilience; and keep regulatory frameworks fit for purpose as AI spreads across financial markets, supporting growth for “Main Street and Wall Street.”
From Constraint to “Non-Adoption Is the Risk”
The most quotable shift is the language. Bessent said Treasury “will continue evaluating regulatory frameworks and enforcement policies” to support the sector’s AI leadership. Deputy Assistant Secretary for FSOC Christina Skinner went further: AI adoption is “critical to America’s financial stability,” and when institutions cannot deploy tools that improve fraud detection, credit allocation, and operational resilience, the system becomes “less efficient and less secure.”
Treasury Chief AI Officer Paras Malik (also Counselor to the Secretary) put the emphasis on operationalization — “embedding AI into core workflows” that measurably improve risk management and resilience — and described the series as “convening regulators and industry leaders to ensure governance frameworks evolve alongside deployment.”
In plain terms: the regulatory instinct of the past decade was “prove the new technology harmless first.” The official position now is that a bank which refuses AI is itself a fresh source of systemic risk.
The Groundwork Already Laid: Lexicon and Risk Framework
The series also does not start from zero. In February 2026, Treasury published an AI Lexicon to standardize key terminology and a Financial Services AI Risk Management Framework, plus six resources covering governance, data practices, transparency, fraud, and digital identity. Per the official series page, the first roundtable already took place on March 4, focused on strategy and governance; the second is set for April 7 on value generation.
So the March 23 announcement upgrades existing paperwork into a standing dialogue mechanism: define the terms first, hand over a framework, then open the table on implementation.
What It Means for Finance Teams and Builders
Three observations. First, for bank and insurance technology teams, the wind has clearly shifted toward “go”: compliance departments now hold an officially sanctioned pro-adoption argument instead of a default no. PYMNTS framed the whole initiative as an attack on regulatory friction — the accumulated interpretation and examination practices that keep banks slow to deploy AI even where the technology is already proven. Second, for fintech developers, the four roundtables are a window into which AI use cases regulators name out loud — the repeatedly cited areas of fraud detection and credit allocation will see demand first. Third, do not ignore the data: actual AI adoption in finance can be cross-checked against measured usage data such as the Anthropic Economic Index, to see where official narrative and real usage diverge.
The risk sits in the same place as the ambition. Defining non-adoption as risk grants institutions political license to deploy widely; whether governance truly keeps pace with deployment is what will decide whether this series succeeds.
Sources
- Treasury Launches the Artificial Intelligence (AI) Innovation Series — U.S. Department of the Treasury
- Treasury Department Targets Regulatory Friction to Scale Bank AI Adoption — PYMNTS
- Artificial Intelligence Innovation Series — U.S. Treasury
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.