AIEOG Shared AI Lexicon
A shared vocabulary for AI in financial services, published February 2026 by the AI Executive Oversight Group — a public-private partnership formed by the US Treasury with FBIIC and FSSCC. Defines roughly seventy terms drawn from standards, academic publications and government sources, including Agentic AI, AI Agent, AI drift, guardrails, human-in-the-loop, model risk, and third-party AI risk.
Why it matters for autonomous finance
This is the sector's official vocabulary, and it matters more than its modest framing suggests. Definitions determine scope: the Lexicon defines an AI Agent as a system that autonomously perceives its environment, decides what to do, and takes actions to achieve its goals — a Treasury-convened body describing exactly the class of system that SR 26-2 then placed outside model risk management two months later. Where a term exists here, use it; divergence should be deliberate and stated.
What it does not cover
Explicitly an optional tool, and expressly not intended for legal interpretation of regulations, supervisory statements, international agreements or private contracts. It does not represent the official views of Treasury or any participant. It is a communication aid, not an authority — and it contains no term for delegated authority, agent commissioning, or the boundary of what an agent may commit.
No vocabulary for authority grants, conformance, or ownership of agent output. Every entry in this register carries this section. A standard read past its scope is worse than no standard, because it produces confidence without coverage.