AI Regulation
Regulation is read here for one thing: what it obliges an operator to produce, and when. Summaries are abundant; the mapping from obligation to evidence is not. These posts stay on that mapping. See the EU AI Act approach, the Digital Omnibus mapping, and what replaced SR 11-7.
SR 26-2 Leaves Generative AI Out of Scope: How Banks Govern LLMs and Agents Anyway
SR 26-2 replaced SR 11-7 and left generative and agentic AI out of scope. How banks govern LLMs and agents anyway, practice by practice, and what to keep.
Third-Party Risk Management for AI Model Providers
A rented AI model is a third-party relationship. How banks apply TPRM to AI model providers, what to ask, and the evidence to keep in their own records.
One Global AI Policy Plane, Including China
How a multinational gives its China teams governed AI on in-country models, under the same identity and policy as the rest of the company, with prompts, responses and records kept in China.
One governed trust layer for the business of promises
Insurers will rent some intelligence and own more of it. Claims, underwriting and service run on both, and every call answers to the same rules and the same examiner.
Japan's Cyber Emergency: Practical Steps for Nations and Enterprises Before the Next Wave
One cloud outage reached 495 organizations at once. A stolen license image could fetch 25 yen. What Japan's response teaches about governing AI traffic in both directions.
FINRA Rule 3110 and AI Agents: A Supervision Guide
Model risk guidance stepped back from agentic AI. FINRA did not. What Rule 3110 requires of a broker-dealer deploying AI agents, and what the supervisory record has to contain.
The CISO's Guide to Runtime Governance for AI Agents [2026]
Runtime governance applies policy to each AI agent call before it executes. What it controls, where it runs, and how it maps to SR 26-2 and FINRA 3110.
The Trust Fabric: A four-layer architecture for governing AI agents
The Trust Fabric, a four-layer architecture for governing AI agents
Six Intelligence Agencies Just Published the Runtime Governance Spec
NSA, CISA, ASD's ACSC, Canadian Cyber Centre, NCSC-NZ, NCSC-UK. Thirty pages. The recommended controls map directly to the runtime plane that workflow platforms do not address.
Smartflow Sovereign: Built for the EU AI Act
The runtime evidence the EU AI Act asks of high-risk systems, due December 2, 2027 for Annex III, and how Smartflow Sovereign records it at the call.
EU AI Act Year One: Prohibited Practices Review and What's Coming Next
The Commission's mandatory review of prohibited AI practices may expand the ban list. Here's what enterprises should monitor.
The Colorado AI Act Takes Effect: First Major US State AI Compliance Deadline
Colorado's comprehensive AI legislation becomes enforceable February 1. Here's what enterprises need to know about the most ambitious US state AI law.