Nearly three-quarters of corporate boards have only moderate or limited expertise in artificial intelligence, according to KPMG’s Global AI Pulse Survey. That finding sits behind a set of AI Board Governance Principles that KPMG International and the INSEAD Corporate Governance Centre published jointly on 14 April 2026 — a sector-agnostic framework aimed less at AI strategy and more at a narrower, harder question: what is a board actually responsible for once AI decisions move faster than board cycles can review them.

The five principles

The framework sets out five areas of board responsibility.

Strategic oversight for long-term value creation in experimental environments. Boards are asked to oversee AI investment and deployment as a strategic capability, not a technology purchase — recognizing that AI initiatives will often run as ongoing experimentation rather than a single, approvable project.

Active technology and security oversight, explicitly framed as balancing sovereignty — retaining control over AI systems and the data they use — against the agility a competitive AI program requires.

Workforce transformation and human accountability. The principle centers on preserving human judgment and accountability as AI takes over more operational decisions, rather than treating productivity gains as self-justifying.

Building trustworthy AI that reflects the organization’s stated values and applicable regulatory requirements — positioning trust as a governance output, not a technical property of the system.

The work of the board itself. The fifth principle turns the lens inward, examining how AI changes board processes, information flow and decision-making — not just what boards oversee, but how they oversee it.

Why this framework, and why now

Annet Aris of INSEAD describes the principles as “sector-agnostic” with “worldwide applicability” across organizations at different levels of AI maturity — language chosen deliberately to make the framework usable by a board just beginning to encounter AI-related decisions and one already running AI in production. Steve Chase of KPMG frames the stakes in governance terms rather than technology terms: “Trust in AI — and in the governance behind it — is what turns ambition into durable value.”

Analysis: the framing matters. This is not an AI-adoption playbook, and it does not attempt to tell boards which AI investments to make. It is a governance-capability framework arriving at a moment when, per KPMG’s own survey, most boards lack the expertise to independently assess AI risk — which means the practical test of the framework is not whether boards adopt its five principles as stated, but whether boards close the expertise gap the survey identifies enough to apply them with judgment rather than by checklist.

The principles also arrive alongside a broader wave of board-level AI governance activity — WilmerHale’s January 2026 client alert on board oversight priorities and ongoing AI governance and control checklists circulating among governance advisors all point the same direction: 2026 is the year AI oversight moved from a topic boards discuss to a competency boards are expected to demonstrate.

What boards are actually being asked to do differently

Reduced to practice, the five principles ask boards to do four things most have not historically done in a formal, recurring way.

Treat AI oversight as continuous, not project-gated. Traditional board oversight approves discrete initiatives at discrete points. AI’s “experimental environment” framing asks boards to oversee an ongoing capability instead — closer to how boards already oversee cybersecurity risk than how they approve capital projects.

Name where human judgment is preserved, explicitly. The workforce-transformation principle is not a call to slow AI deployment; it is a call to document, deliberately, which decisions keep a human accountable and which do not — a distinction most organizations currently make implicitly, if at all.

Close the expertise gap directly, not by delegation. With close to 75% of boards reporting only moderate or limited AI expertise, the honest first step for most boards is a capability assessment of the board itself, not another AI strategy briefing from management. Independent advisory support — brought in specifically to assess board-level AI governance readiness rather than to advise on AI strategy — addresses a different gap than most boards currently resource for.

Audit the board’s own information flow. The fifth principle’s inward focus is easy to skip past in favor of the more visible first four. It is also the one most directly actionable without any AI deployment decision at all: reviewing how AI-related information currently reaches the board, and whether that flow is adequate, can start immediately.

Frequently asked questions

What are the AI Board Governance Principles?

A five-principle framework published jointly by KPMG International and the INSEAD Corporate Governance Centre on 14 April 2026, covering strategic oversight, technology and security oversight, workforce transformation and human accountability, building trustworthy AI, and the board’s own AI-era processes.

How much AI expertise do boards actually have?

According to KPMG’s Global AI Pulse Survey, nearly 75% of boards report only moderate or limited AI expertise.

Is this framework specific to one industry?

No. It is described by its authors as sector-agnostic, intended to apply across organizations at different stages of AI maturity.

Do the principles tell boards which AI investments to approve?

No. The framework addresses governance responsibility and oversight capability — how boards should oversee AI decisions and processes — rather than making specific technology or investment recommendations.

Sources

Principles, launch date and quotations from the KPMG International and INSEAD Corporate Governance Centre press release, 14 April 2026, and KPMG’s Global AI Pulse Survey as cited therein. Related governance context from WilmerHale’s January 2026 client alert. Boards evaluating their own AI governance posture should consult the full principles document and qualified governance counsel.

PGAN advises boards and executive leadership on AI governance readiness and technology oversight capability. Explore our AI & technology advisory and governance advisory practices, or request advisory. Read our related briefing on construction’s AI pilot-to-production gap.