Governing Agentic AI

AI tools are already in use across municipal and school district workflows, often informally and without a governance framework, procurement policy, or audit trail. Agentic AI, systems that take actions on an organization's behalf rather than simply answering questions, is arriving even faster, and it raises the stakes: an AI agent with access to systems and data is a new kind of attack surface if it isn't governed properly. This session offers Connecticut's public sector and education leaders a practical, non-technical framework for evaluating and governing AI adoption: what questions to ask vendors, what an audit trail should actually capture, and how to get the productivity benefits of AI without introducing new risk into networks that already carry significant public trust. Drawn from direct experience advising public sector organizations on responsible AI governance.
Most AI governance conversations aimed at the public sector focus on high-level ethics principles or federal policy, neither of which gives a first selectman, superintendent, or town IT director something they can act on Monday morning. This session is built for that gap. It opens with why agentic AI (AI that can take actions such as sending communications, modifying records, or interacting with other systems) is fundamentally different from chatbot-style AI in terms of risk surface, and why that distinction matters for procurement and oversight decisions being made right now. It then walks through a practical governance framework covering vendor evaluation questions, the components of a defensible audit trail, human-in-the-loop requirements for higher-risk actions, and how to structure a pilot so it stays contained. The session draws on direct, current experience advising Connecticut public sector and education organizations on responsible AI adoption, and is designed to leave attendees with language and criteria they can bring directly into their own procurement and IT governance discussions.

PDF - Governing the Agents - Oct 2026 (pdf)Download

PDF - Take away - AI Governance Toolkit (pdf)Download

AI as Weapon and Shield

Artificial intelligence has become a tool for attackers and defenders alike, and the gap between the two is closing fast. This session breaks down how threat actors are already using AI in the field: deepfake audio and video used in CEO and vendor impersonation fraud, AI-generated phishing that defeats the grammar and formatting cues staff are trained to spot, and voice-cloning scams targeting finance and HR teams. It then turns to the defensive side, showing how AI-powered anomaly detection, automated triage, and threat correlation are changing what a modern SOC can catch and how fast. Attendees will leave with a practical framework for assessing where their organization is exposed to AI-enabled attacks today, and where AI-assisted defense tools can realistically help, without the hype.
Connecticut's municipalities, school districts, and state agencies face a threat landscape that has changed faster than most security awareness programs have kept up with. This session presents a grounded, non-vendor-driven look at how generative and agentic AI are reshaping both sides of the cybersecurity equation for community anchor institutions specifically: small IT teams, limited budgets, and high public trust stakes. The first half covers offensive use cases with real-world patterns: AI-assisted social engineering, deepfake-enabled fraud against finance and procurement workflows, and AI-generated malware and phishing infrastructure. The second half covers what AI-assisted defense actually looks like in practice for resource-constrained organizations, distinguishing genuine capability from marketing. The session closes with a practical checklist attendees can bring back to their own IT and leadership teams. This complements threat briefings by giving attendees a specific, actionable lens on the AI dimension of current and emerging threats.

PDF - AI as Weapon and Shield - Oct 2026 (pdf)Download

PDF - Take Away - AI Threat Readiness Checklist (pdf)Download

Governing Agentic AI

AI as Weapon and Shield

This website uses cookies.