Board Committees and AI: Should You Create a Dedicated AI Oversight Committee?

The arguments in favour of a dedicated AI oversight committee centre on the complexity, pace of change, and cross-functional nature of AI governance.

AI governance cuts across the responsibilities of every existing board committee. It has technology risk dimensions that sit with audit. It has workforce implications that sit with nomination. It has performance dimensions that sit with the main board. In practice, this means AI governance falls between committees rather than being owned by any of them.

A dedicated AI committee creates a home for AI governance at board level, with the mandate to develop expertise, set standards, and provide oversight that crosses functional boundaries. Companies that have created such committees, including several FTSE 100 companies that moved early on AI governance, report that they have improved both the quality of board AI discussions and the consistency of AI governance across the business.

The arguments against a dedicated AI committee are largely structural. Committees work well when they have stable, defined subject matter, a reasonable cadence of substantive decisions to make, and members with relevant expertise. AI governance in 2025 may not yet meet all three criteria for every organisation.

For many boards, the frequency of material AI governance decisions is not yet high enough to justify standing committee meeting time. The expertise requirements for effective AI committee membership are specialised in ways that existing director pools may not meet. And the risk of creating a committee that provides governance theatre rather than governance substance is real.

An AI committee that meets quarterly, receives the same briefing that the full board would have received, and makes no decisions that the full board could not have made is not adding governance value. It is adding governance cost and potentially creating the illusion of oversight where substantive oversight is lacking.

The most common effective approach among early-adopting UK companies is not a dedicated AI committee but enhanced AI oversight within existing committee structures.

The audit committee takes responsibility for AI risk oversight: reviewing the AI risk register, assessing the adequacy of AI controls, and ensuring that internal audit covers AI governance as a risk area. This is a natural fit with audit's existing risk oversight mandate.

The main board takes AI strategy as a standing agenda item, receiving updates on AI programme performance, competitive AI position, and material AI investment decisions on a quarterly basis. This ensures that AI strategy remains a full board responsibility rather than being delegated to a subcommittee that the full board then rubber-stamps.

A dedicated AI working group at executive level, reporting to the board, provides the operational AI governance function that neither the audit committee nor the main board is well-suited to perform: policy development, deployment oversight, incident investigation, and regulatory monitoring.

A dedicated AI oversight committee is most justified when the board has at least two or three directors with sufficient AI expertise to make committee discussions substantive, when the frequency and materiality of AI governance decisions justifies standing meeting time, and when the organisation's AI programme is large enough that a committee focused specifically on AI can identify governance issues that would otherwise be missed.

For most FTSE 250 companies in 2025, the threshold for a dedicated committee is probably not yet met. For FTSE 100 companies with large AI programmes and multiple directors with AI expertise, the case is stronger. The decision should be made on the basis of governance substance, not on the basis of demonstrating to shareholders that AI is being taken seriously.