The Board Has a New Question: Can We Trust Our AI?
Non-executive directors are increasingly asking management teams to explain AI governance. Most leadership teams are not prepared.
Something shifted in boardrooms in late 2024. Non-executive directors — particularly those with regulatory, legal, or risk backgrounds — started asking pointed questions about AI governance that management teams were not prepared to answer. Not 'what AI are we using?' but 'how do we know our AI is doing what we think it's doing?'
The regulatory signal was clear: ESMA, EBA, and national competent authorities across the EU are treating AI governance as a board-level responsibility, not an IT function. This means the audit committee needs to understand model risk. The risk committee needs to understand AI failure modes. The remuneration committee needs to consider whether AI-driven performance metrics create perverse incentives.
A useful framework for board-level AI governance has three layers: transparency (do we know what AI systems we have and what decisions they influence?), accountability (who is responsible when an AI system fails?), and controllability (can we audit, adjust, or shut down AI systems when needed?). Most organisations score well on the first layer and poorly on the second and third.