Glass Lewis has documented, with survey data and S&P 100 proxy analysis, that 72 percent of the largest US companies have not met the governance standard that most institutional investors now expect: board-level AI oversight disclosure backed by a formal AI policy. The liability vector is no longer theoretical. Proxy advisors are scoring this in the current season. Boards that have approved major AI investments without codifying oversight in committee charters are the ones exposed.
PRIORITY 9 | SILO: INSTITUTIONAL
Glass Lewis’s S&P 100 analysis finds only 28% of the largest US companies disclose both board-level AI oversight and a formal AI policy, while 65% of institutional investors now expect that disclosure. The 2026 proxy season is active.
The Signal
Glass Lewis, in research published on the Harvard Law School Forum on Corporate Governance on March 11, 2026, documented the AI governance gap across the S&P 100 with primary survey data and direct proxy statement analysis. Of the 100 largest US companies by market weight, 54% disclosed some form of board-level AI oversight in their 2025 proxy statements. Only 28% disclosed both board oversight and a formal AI policy.
The separate Debevoise & Plimpton analysis published April 27, 2026 on the same forum makes the fiduciary frame explicit: the SEC has already made clear regarding cybersecurity that boards must find a way to exercise supervisory obligations even in technical areas when those areas present enterprise risk. AI now meets that threshold for most Fortune 500 companies.
The Evidence
Glass Lewis’s 2024 and 2025 policy surveys establish the investor expectation with specificity: 65% of US investors expect all companies to provide clear disclosure of the board’s oversight of AI governance and AI ethics. 49% say that oversight should be codified in a committee charter or governing document. 46% say a standing committee, not ad hoc management, should carry the responsibility.
The S&P 100 proxy analysis covers January through December 2025 filings. Of companies disclosing board oversight (54%), 63% routed it to a specific committee, most often audit or technology. The remaining 37% assigned it to the full board without structural designation, a posture Glass Lewis flags as less defensible under investor scrutiny.
The SEC’s Investor Advisory Committee, in a December 2024 recommendation, directed issuers to disclose board oversight mechanisms for AI deployment and to report material effects of AI on internal operations. That recommendation is not a rule, but it signals the direction of future SEC guidance and shapes the expectations institutional investors are applying now.
The Strategic Implication
Defensive Risk. Audit committee chairs at S&P 100 companies that have approved material AI investments but have not codified oversight in their committee charters are the specific exposure point. The mechanism: Glass Lewis evaluates AI oversight on a case-by-case basis for 67% of its institutional clients, meaning an adverse recommendation against a governance committee member is available without a new policy trigger. The timing is the current proxy season, before June annual meetings when most S&P 100 companies take shareholder votes. The responsible defense move is to confirm, in writing, that the audit or risk committee charter explicitly names AI risk as a standing oversight responsibility, and to ensure that 2026 proxy disclosures reflect both oversight structure and a summary of the AI governance framework.
Offensive Advantage. Boards that move now to codify AI oversight, publish the committee designation, and disclose the governance framework in supplemental proxy materials are positioned to separate themselves visibly from the 72% majority that has not done so. Lockheed Martin’s approach, identified by Glass Lewis as the exemplar among the S&P 100 companies studied in depth, produced a measurable result: no AI-related shareholder proposals appeared on the company’s 2025 ballot, directly attributable to proactive governance disclosure and investor engagement. The board that requires a disclosed AI governance framework before approving the next major AI investment is the board that avoids the activist-vector opening that 72% of the S&P 100 has left exposed.