---
title: "AI Governance: The Board Disclosure Gap | Third Economy"
description: Explore what investors expect from board AI oversight, from clear accountability and risk inventories to provider reviews, incident response and disclosure.
image: https://pages.thirdeconomy.com/hubfs/Thought%20Leadership%20Webpage%20Graphic-1.png
---

[![Third-Economy-logo-horizontal-1](https://pages.thirdeconomy.com/hs-fs/hubfs/Third-Economy-logo-horizontal-1.png?width=1500&height=381&name=Third-Economy-logo-horizontal-1.png "Third-Economy-logo-horizontal-1")](https://www.thirdeconomy.com/)

[SUSTAINABILITY SERVICES](https://www.thirdeconomy.com/corporate-services/)

[INVESTOR SERVICES](https://www.thirdeconomy.com/investor-services/)

[VIA3](https://www.thirdeconomy.com/ratings/)

[ABOUT](https://www.thirdeconomy.com/about/)

[NEWS & EVENTS](https://www.thirdeconomy.com/news-events/)

[CONTACT](https://www.thirdeconomy.com/contact/)

![Thought Leadership Webpage Graphic-1](https://pages.thirdeconomy.com/hubfs/Thought%20Leadership%20Webpage%20Graphic-1.png "Thought Leadership Webpage Graphic-1")

# AI governance: The next disclosure gap boards have to close

***By Third Economy, November 2026***

Companies are deploying AI faster than they are governing it. The tools reach hiring, pricing, product safety and customer data before a board process exists to review them, and oversight is proving harder to show than adoption. That gap between deployment and governance is where investor questions now land, and no final rulebook is required for your board to be asked about it.

The disclosure record shows the gap. 83% of S&P 500 companies disclose AI as a material risk, yet only 8% disclose board-level oversight of AI, and 9% have a written AI policy (ISS, 3,048 U.S. companies, 2026). Nearly every large company calls the risk material, almost no one can show what the board does about it. Directors rate their own AI expertise last among the capabilities surveyed (Diligent), and an MSCI investor roundtable in spring 2026 concluded that board AI governance is largely symbolic rather than substantive.

The pressure is not waiting for regulation to catch up. The EU AI Act entered its main application phase on August 2, 2026, with selected high-risk requirements phased through 2028 ([European Commission AI Act Service Desk](https://ai-act-service-desk.ec.europa.eu/en/ai-act/timeline/timeline-implementation-eu-ai-act)), while raters moved faster: S&P Global CSA now scores Responsible AI in its Governance dimension, ISS QualityScore added AI oversight factors in October 2025, and Glass Lewis finds 65% of investors want disclosure of board AI oversight. Stewardship leaders representing $55 trillion in assets under management told EY they want to understand board AI fluency and oversight ([EY Center for Board Matters, 2026](https://www.ey.com/content/dam/ey-unified-site/ey-com/en-us/campaigns/board-matters/documents/ey-cbm-five-ways-boards-can-respond-to-investor-expectations-on-ai.pdf)). When investors and raters at that scale add AI to their agendas, the questions reach your board whether or not your disclosure is ready.

## AI risk oversight belongs to the board

AI decisions now shape strategy, operations, customer relationships and capital allocation. When a tool influences hiring, product safety or sensitive data, its consequences reach well beyond the technology team. You do not need to review every use case. You do need confidence that management can identify material risks and respond when controls fail.

There is no one-size-fits-all committee model. Some boards assign AI to an existing risk, audit or technology committee, while others retain it at the full board. Investors care more about clear accountability, relevant director expertise and the quality of information reaching directors than about any prescribed structure.

## The questions investors will ask you

Investor questions are practical, and the list is short enough to prepare for. Which committee owns AI risk? What reaches the board, how often and from whom? If an AI incident occurs, who escalates it and who decides the response?

A diligence process built from the investor’s side of the table shows how testing targets methodology rather than messaging, and that is the vantage point our team brings to this work. Investors want a working view of your AI footprint: material uses, business purpose, data involved, risk classification and the accountable executive. They ask how you evaluate external providers on reliability, privacy, security and intellectual property exposure, and whether contracts and contingency plans match those dependencies.

The harder question is the one raters now score. “The board endorsed our AI policy” is a scored, dated, public claim made annually, and the follow-up is who produced the assessment behind it and whether that person reports through the function being reviewed. Oversight folded into a risk-topic list, with no decision rights attached, satisfies structure but not evidence. Liability follows inattention that cannot be disproved, so the record of what the board asked, received and acted on is the defense.

They also want the value case. What outcomes is management tracking, and how does it test whether expected productivity or revenue gains are materializing? Adoption alone does not prove a return. This is corporate value, not values: investors are testing whether oversight protects the business, not whether your language sounds responsible.

## A framework that can adapt

Regulation is developing unevenly. The EU AI Act applies in phases, Colorado’s AI law and other state measures add requirements in the United States, and sector guidance applies in areas such as healthcare, finance and employment ([Colorado SB25B-004](https://leg.colorado.gov/laws/session-laws/SB25B-004/3/download)). The details and timelines can change, so map obligations to your activities and markets, then revisit the analysis as guidance evolves.

A governance foundation built once can deploy everywhere. Define decision rights, set escalation thresholds, assess material use cases, oversee external providers and monitor outcomes. Then connect AI review to your existing processes for privacy, cybersecurity and operational resilience. That reduces duplicated work and helps directors focus on the risks that matter to your business, the same governance and risk management discipline investors are prioritizing to mitigate volatility ([Third Economy, 2025 Insights Trends](https://pages.thirdeconomy.com/2025-insights-trends?hsLang=en)).

Colorado’s repeal of its AI act before it took effect is the caution against building to a rulebook. The duty the courts recognize already exists: directors answer for a sustained failure to maintain and monitor a reporting system on mission-critical risk, and the Marriott cyber litigation extended that reasoning to technology oversight. The record, not the statute, is what stands behind you.

## Connect oversight to your existing reporting

AI governance should strengthen the reporting you already produce, not create a disconnected program. Where relevant, connect it to proxy materials, annual reports, enterprise risk reporting and sustainability disclosures so investors see how oversight, controls and strategy fit together.

A credible first disclosure is modest and specific. Identify the board or committee responsible, describe management accountability, explain how directors receive updates, and summarize your approach to inventorying material uses, reviewing provider risk and handling incidents. Plain, measured disclosure beats broad claims about responsible AI that carry no evidence, and charter language with a decision right attached beats a risk-topic list every time.

## Questions we hear most

### Who should own AI risk at the board level?

Investors expect a clearly stated assignment, coordination across committees and full board visibility when AI presents material strategic or financial risk.

### What belongs in an AI use inventory?

Material internal and external uses, their business purpose, affected stakeholders, relevant data, accountable owners and risk controls, including how you review vendor changes and incidents.

### Do we have to publish technical details?

No. Useful disclosure explains oversight, material risks and controls at a level investors and raters can assess without revealing confidential information.

### What evidence should the board keep?

The record of what was asked, what was received, who produced it and whether that reporting line is independent of the function reviewed. A stated policy alone never scores.

## In short

AI governance is fast becoming a measure of board effectiveness because it links strategy, operational risk and access to capital. Waiting for a final rulebook is not a risk management strategy. Begin with a board-approved owner, a current inventory of material AI uses and a documented incident and provider review process, then connect those elements to your existing risk and disclosure systems so evidence builds as you deploy. We help boards set these priorities through our [Public-Market Readiness](https://www.thirdeconomy.com/corporate-services/) work.

If we can be helpful as you navigate AI governance and investor disclosure, please don’t hesitate to [reach out](https://www.thirdeconomy.com/contact/).

*The information provided does not, and is not intended to, constitute legal advice; instead, all information, content, and materials available are for general informational purposes only.*

[Discuss AI governance](https://www.thirdeconomy.com/contact/)

### Want to receive our latest news and events?

[Sign up here](https://www.thirdeconomy.com/contact/)

---

![Logo-1](https://pages.thirdeconomy.com/hs-fs/hubfs/Logo-1.png?width=200&height=47&name=Logo-1.png "Logo-1")

#### CONTACT

## Third Economy is a sustainable investment research and consulting firm. We work with corporations and institutional investors to understand how sustainability factors impact financial value. Our clients leverage this understanding to enhance and promote their own capabilities, thus building a more sustainable economy, a third economy.

<https://www.linkedin.com/company/thirdeconomy> <http://www.facebook.com/thirdeconomy> <http://www.thirdeconomy.com/> <https://www.thirdeconomy.com/contact/>

  370 7th St Suite 9, San Francisco, CA 94103

  +1 415 935-0030

  [Click here for business inquiries](https://www.thirdeconomy.com/contact/)

  [info@thirdeconomy.com (all other inquiries)](mailto:info@thirdeconomy.com)

---

2024 Third Economy, Inc.  
All rights Reserved.

[SUSTAINABILITY POLICY](https://www.thirdeconomy.com/wp-content/uploads/2023/09/3ECON-sustainability-policy.pdf)

[DIVERSITY POLICY](https://www.thirdeconomy.com/wp-content/uploads/2023/09/3ECON-diversity-policy.pdf)

[TERMS AND CONDITIONS](https://www.thirdeconomy.com/terms-and-conditions/)

[PRIVACY POLICY](https://www.thirdeconomy.com/privacy-policy/)

[COPYRIGHT POLICY](https://www.thirdeconomy.com/copyright-policy/)

[SITEMAP](https://www.thirdeconomy.com/sitemap/)