Railpen, the UK rail workers’ pension fund with £36bn (€40.5bn) in assets, has found that companies’ AI governance and risk-management practices lag behind their adoption of the technology, following an assessment of a portion of its active equity holdings.
In partnership with Chronos Sustainability, Railpen has published a report – Navigating AI: An investor framework for assessing exposure and governance – introducing a benchmarking tool based on its AI Governance Framework (AIGF).
The report aims to help investors assess corporate AI exposure, benchmark governance maturity, evaluate governance practices and identify areas for engagement.
Building on Railpen’s 2025 report, Achieving effective AI governance, the latest research translates the AIGF into a benchmarking tool for assessing governance practices across portfolio companies.
“We are aware that, as AI companies begin to provide the foundational underpinning for a lot of the economic shifts that we’re going to see in the future, the kinds of governance issues that tend to be concentrated at founder-led tech-heavy firms are going to be seen as models for the governance arrangements at companies even beyond the tech sector,” said Caroline Escott, co-head of sustainable ownership and head of investment stewardship at Railpen.
Railpen and Chronos Sustainability applied the AIGF to around 30 listed companies in Railpen’s active equity holdings. The companies were assessed against 18 questions across three pillars: governance and strategy, risk management, and performance reporting.
Findings
The assessment found that AI governance is uneven and frequently underdeveloped relative to companies’ exposure to the technology.
Around half of the companies assessed referred to responsible AI principles in a policy, but few explained how those principles were implemented, monitored or enforced, according to the report.
Speaking about the findings, Jasmine Porter, investment analyst in Railpen’s sustainable ownership team, said: “As a universal owner, you can’t simply diversify away from systemic risk. Instead, you need to understand where those risks and opportunities lie, and then be thoughtful about where you can have the greatest influence.”
Railpen said the findings could help it identify companies with high AI exposure and weak governance, prioritise stewardship resources and incorporate AI risks into investment due diligence.
