About Michael Watson

Front-facing headshot of Michael Watson in a white collared shirt
Headshot: 2026 Cybersecurity Modernization Summit.
Technology Leadership in Public Service

As Chief Information Officer for the Commonwealth of Virginia, my work sits at the intersection of public service, cybersecurity, infrastructure, governance, and emerging technology.

The responsibility is to make complex systems understandable enough to govern and reliable enough to use.

Professional conversations around technology, research, and speaking are welcome.

View selected speaking engagements.

Explore interviews and media · Read about my CIO appointment

Connect on LinkedIn
Evidence, Boundaries, and Human Judgment

My research interests include AI capability and uncertainty, autonomous cyber threats, trustworthy communication, and agent systems with meaningful controls.

When should a system answer, ask for more evidence, or hand a decision to a person? How can an evaluation distinguish repeatable capability from a result that depends on a particular prompt? What makes an automated action explainable after it has happened?

These questions connect research to practical leadership. They shape how systems are tested, how confidence is communicated, and how responsibility is assigned when automation reaches the edge of what it can support.

Education and Applied Research

MSE, Telecommunications with a focus on Security and Management, University of Pennsylvania.

BS, Computer Science with a Networking concentration, James Madison University.

Current research themes include AI capability boundaries, cybersecurity model evaluation, representation reliability, and human escalation.

View selected work
Academic Degrees
Featured Essays
Areas of Focus
Research Themes

Selected Recognition

Writing and Ideas

Published essays and the principles that connect the work.

A hand writing with a pen beside a notebook and mug

Teaching is a feedback loop toward mastery. AI learning systems should test explanation and application, not simply track completion.

The AI Ate My Homework and Then Asked Me to Explain It

Blue and gray network cables connected to a patch panel

Autonomous attacks reveal themselves through relationships and behavior across systems, not isolated alerts.

When Attacks Become Autonomous, Defenses Must Become Insightful

A magnifying glass, notebook, and pencil on a pale desk

When AI can imitate an organization’s voice, trust requires visible endorsement, evidence, accuracy, and authoritative anchors.

The Truth Is Out There… and We Can Prove It

Library bookshelves photographed with shallow depth of field

Good decisions make their supporting evidence visible and state the limits of what that evidence supports.

Evidence Before Confidence

Hands holding a map and compass outdoors

Automation should make responsibility and escalation clear. Systems need a useful handoff when the evidence is insufficient.

Meaningful Human Judgment