Investor Expectations and Company Disclosures on Governance of AI

Investor Expectations Are Outpacing Boardroom Disclosure on AI Governance

Why listed companies need to demonstrate effective board oversight of artificial intelligence

Artificial intelligence has moved from an emerging theme to a core item on the investor stewardship agenda. Our review of voting guidelines, position papers, and engagement practices at 65 major investment managers finds that AI governance has become a defined area of scrutiny in a remarkably short period: 26 firms now set out explicit expectations on AI governance in their voting guidelines or position papers, and a further 15 describe AI-related engagement activity in their stewardship reporting. Shareholder proposals on the topic, while not yet successful, have also attracted nearly double the support levels seen for comparable ESG proposals over the same period.

Investors are converging around a consistent set of expectations. They want to see boards with the skills and structures to oversee AI meaningfully; a formal, board-endorsed policy framework covering AI ethics and responsible use; explicit consideration of AI within strategy and enterprise risk management, extending well beyond cybersecurity; and disclosure that connects all of this together across the strategy, risk, governance, and sustainability sections of the annual report.

Our parallel review of disclosure practices at the 30 largest FTSE-listed companies by market capitalization suggests that boards are still building toward these expectations. Only four of the 30 companies reviewed offered any reflection on directors' AI-relevant skills and expertise, and just 13 referenced board-level training in AI. Two-thirds of companies reported some board-level discussion of AI's strategic implications, and a similar proportion pointed to committee-level involvement, most commonly through existing audit or risk committees rather than a dedicated technology committee. Disclosure on AI policy frameworks is also emerging unevenly: roughly six in ten companies indicated that a policy was in place, but few provided the level of detail needed for investors to assess its robustness. Coverage of AI's broader social and environmental dimensions, beyond data privacy, remains particularly limited.

The direction of travel is clear. Boards should expect AI governance to feature more prominently in engagement and voting decisions in the seasons ahead, and should treat the gap between current disclosure and investor expectations as an opportunity to get ahead of the issue rather than a compliance afterthought. Practical starting points include revisiting how director skills and AI-related training are described in nomination committee reporting, clarifying where accountability for AI oversight sits within the board and its committees, and ensuring the annual report tells a connected story about AI across strategy, risk, and governance rather than treating it as a single, isolated disclosure.

Download our one-page overview for a snapshot of these developments and additional insights. If you are interested in our full report or you would like to discuss in further detail, please contact us.