One oversight signal. One governance insight. One action item.

Once a month. Read in under 3 minutes.

Center Papers

Center Paper
The AI Oversight Governance Layer

Three layers of governance operate in any organization deploying AI. The oversight layer, which decides what is governed, at what granularity, and who is accountable, is usually missing.

Center Paper
Cognitive Sovereignty and the Capacity to Decide

An informed decision requires a decision-maker still capable of making it. As institutions embed AI into consequential decisions, the board must govern the human judgment around the system, not only the system.

Center Paper
Agentic AI and the Oversight Lens

When AI acts rather than advises, model risk standards fail. Why the oversight program is the only standard that reaches it.

Center Paper
Builders and Buyers

Accountability does not transfer with the vendor contract. Four bodies of law, four industries, and one unifying principle.

Center Paper
The Committee Fallacy

Assigning AI risk to a committee is not governing it. Why the Delaware doctrine demands a program, not just a venue.

Center Paper
Regulatory Enforcement Landscape

Mapping the 2026 enforcement environment. Four forces converging on one demand, despite repeals and preemption fights.

Center Paper
The Caremark Liability Roadmap

How the duty of oversight applies to AI, and the legal standard for insulating the board from bad faith failure-to-monitor claims.

Center Paper
AI Oversight Baseline Review

A free fourteen-domain self-review of the oversight layer, online or as a PDF.

Center Paper
The Two Duties

Examining the dual responsibilities of AI oversight: the duty to oversee the system's compliance and risk, and the duty to oversee the human capacity to direct it.

Reference Tool
Who Decides, Who Answers: A RACI for AI Oversight

Decision rights and accountability across the eight domains of an AI oversight program. The board and its committees own the WHAT and the WHO. Management owns the HOW.

The AI Oversight Brief

June 2026
Deadlines move. The obligation does not.

The EU's Digital Omnibus package is provisionally agreed but not adopted. Why institutions planning to the expected deferral are making a bet without a record.

Essays

The Test
The Informed Decision Standard

The honest answer is a test, not a checklist: can the organization demonstrate, in evidence, that it made an informed decision about AI risk?

Brief
Caremark comes for AI

AI oversight is not a new legal theory. It is the existing duty of board oversight applied to the newest mission-critical risk.

Brief
An AI committee is not a governance program

Forming a committee is the most common response to AI risk, and the most misleading. Governance is a program.

The Problem
The Velocity Gap

AI is being adopted faster than most organizations can govern it. The distance between the two speeds is where fiduciary risk lives.

One oversight signal. One governance insight. One action item.

Once a month. Read in under 3 minutes.