At a glance: five key messages
- AI literacy is now a baseline competency for all directors — not a specialism but a threshold skill, equivalent to financial literacy.
- The skills matrix needs updating. Most current frameworks were built for a different era and do not capture AI governance capability.
- The ‘AI expert’ trap: Whole-board literacy, supported by management and outside advisers, is more effective and better for board dynamics.
- Nomination committees need to ask different questions. AI governance competence should feature in every succession conversation, not as an afterthought.
- Regulatory pressure is real. The 2024 FRC Code, the EU AI Act, and growing investor expectations mean boards that are not acting are already behind.
Boards have been asking the same question for years: how do we ensure we have the right people around the table? The mechanics of the answer have not changed much — skills mapping, planned refreshment, a nomination committee that earns its keep. But AI has shifted what “the right people” actually means, and the question now carries more weight than most boards recognise.
The evidence is striking. The Corporate Governance Institute, a Dublin-based governance training organisation, in their report “Boardroom Resilience in 2026” found that 85% of directors report feeling confident in their board’s overall effectiveness, yet only 35% maintain that confidence when confronted with specific questions about AI adoption and governance. The research, conducted by independent UK market research firm Censuswide across 500 board directors and C-suite executives in the UK and Republic of Ireland, drew its sample from financial services, legal, healthcare and education. The gap between general confidence and specific competence is precisely where succession planning needs to do its work.
From specialism to threshold competency
For most of the past decade, digital or technology expertise on boards has been treated as a welcome addition: useful if you can find it, rarely a prerequisite. The analogy to financial literacy is instructive. A generation ago, not every director was expected to read a set of accounts with any depth. Then the regulatory environment, governance codes, and investor expectations collectively shifted, and financial literacy became a baseline requirement for every board member — not a specialism held by the finance director and one or two others.
AI literacy is on the same trajectory, and the timeline is shorter. The EY Center for Board Matters, drawing on analysis of Fortune 100 company disclosures found that 44% of Fortune 100 companies cited AI in their descriptions of director qualifications in its October 2025 report, up from 26% in 2024. (“Cyber and AI oversight disclosures: what companies shared in 2025.”). EY spoke with 19 stewardship leaders from institutional investors representing $45 trillion in assets, and found that over the same period, the proportion of institutional investors making responsible AI a priority for board engagement nearly doubled, from 19% to 36%. These are not marginal shifts. They signal a change in what investors expect boards to look like.
Writing in the Harvard Law School Forum on Corporate Governance in February 2026, drawing on EY analysis, the authors concluded that all directors should now possess baseline AI literacy as a minimum, and that companies should consider disclosing which directors have AI-related expertise or experience overseeing AI initiatives. (“How Boards can Lead in a World remade by AI”)
This is not about expecting every non-executive to become a data scientist. It is about being able to ask the right questions, interrogate management’s assumptions, and recognise when the organisation’s exposure to AI-related risk is being understated.
The ‘AI expert’ trap
The tempting response — and one that many nomination committees are actively considering — is to appoint a dedicated AI expert to the board. It feels like a solution. In practice, it often creates a new problem.
An article in the Harvard Law School Forum on Corporate Governance in April 2026, set out some of the risks. The pool of individuals with both deep AI expertise and the qualifications to serve effectively as a public company director is genuinely limited. More significantly, a designated expert can undermine the collective dynamics that make boards work well. Other directors defer.
Challenge diminishes. Over time, the rest of the board has less incentive to build their own understanding, because they can always pass the difficult question to the person who “knows about AI.” (Board Oversight of AI: Do Boards need AI Experts”)
There is also the question of conflicts of interest. Individuals with deep AI expertise frequently carry extensive commercial relationships in the technology sector — investments, advisory roles, vendor ties — that require careful management on a public company board.
The EY Center for Board Matters suggested an alternative: an advisory council can attract the perspectives of younger, tech-savvy professionals who may be less interested in becoming full-time board members, and this can be a practical way to bring AI insight into the boardroom without the governance complications of a full appointment.
The more durable answer is whole-board capability, supported by expert input from management, specialist advisers, and structured education. An AI expert director may be warranted in certain organisations where AI is genuinely central to the business model. But it is not a substitute for systemic board literacy, and it should not be treated as one.
Succession planning that reflects the moment
The skills matrix is the nomination committee’s primary tool, and most of them need updating. AI governance competence needs to be included in the skills matrix as a standard criterion, assessed against every existing director and every potential recruit. The Harvard Law Forum’s April 2026 analysis of board skills went further, arguing that succession planning should specifically target candidates with expertise in AI governance — not just technology broadly — alongside geopolitical risk and other emergent exposures.
What that looks like in practice is a different kind of conversation in the nomination committee. For every succession, the committee should ask: does this candidate understand AI well enough to challenge management? Can they read an AI risk register with the same confidence they bring to a financial statement? If the honest answer is no, that is relevant to the appointment decision — and should be recorded as such. That is not setting the bar too high. It is a reasonable expectation of a director in 2026, and it should apply consistently — not just to technology-facing roles, but to every board appointment.
The regulatory backdrop
This is no longer a matter of aspiration. The FRC’s 2024 UK Corporate Governance Code places the chair’s responsibility for board effectiveness squarely in the frame — and effectiveness, in 2026, includes the capacity to govern AI-related risk. The Code has structurally consolidated accountability at the top. Most notably, Provision 29 — the stringent requirement for the board to make an explicit annual declaration on the effectiveness of all material internal controls (financial, operational, compliance, and reporting) — has entered its live implementation cycle as of January 2026. This has placed ultimate oversight firmly on the board and its leadership. Under the 2024 Code’s definition of “material operational and reporting controls,” managing tech-driven risks is no longer a niche IT issue. A board’s effectiveness is directly judged by its capacity to oversee the strategic implementation and risk mitigation of AI. The FRC publication “Generative and Agentic AI Guidance”, issued in March 2026, is directed initially at audit firms, but signals a direction of travel that all boards should note. The EU AI Act’s Article 4 literacy provision, in force since February 2025, requires organisations to ensure sufficient AI literacy among anyone operating or overseeing AI systems on their behalf. For FTSE companies with European operations, that obligation already applies. Even for those without, it sets a standard that UK regulators are watching and that institutional investors are beginning to reference in their voting and engagement policies — as the EY data on investor priorities confirms.
The IoD’s NEDs Reimagined Commission, reporting in January 2026, identified digital and technical literacy gaps amongst the critical challenges now facing non-executive directors, alongside the need to reframe independence and harness the role of AI in boardroom decision-making. This sits at the centre of current governance thinking, not at the margins.
The chair’s responsibility — and five things to do now
Chapter 2 of The FRC’s Guidance on Board Effectiveness (July 2018) sets out the chair’s role as leading the board, managing the agenda, and ensuring the board functions well as a collective. It identifies the chair as responsible for creating the conditions in which open challenge and honest debate take place within the boardroom. That responsibility now has to include AI literacy. If the chair treats AI capability as someone else’s problem — a technology committee question, or something for the one director who “knows about this stuff” — the rest of the board will follow that lead. The chair’s role is to make AI a normal part of the board’s conversation: a lens applied to strategy, risk, and succession in the ordinary course of business, not a crisis to be managed in isolation.
That requires an honest assessment of where the board currently stands. For most FTSE boards, that assessment will reveal more ground to cover than the board’s self-perception suggests. It also means thinking now about whether the current composition is adequate for the decisions the board will face in the next three years, rather than waiting for a convenient vacancy to act.
With that in mind, here are five things a chair should do:
- Commission an audit of your board’s current AI literacy — not a self-assessment, but a structured review against a clear external benchmark.
- Update the skills matrix to include AI governance as a standard criterion alongside financial and operational expertise, and apply it to all succession decisions from now.
- Brief the nomination committee: every candidate discussion should include an explicit assessment of AI capability, regardless of the role being filled.
- Commission structured AI education for the whole board — not one-off briefings, but a programme that builds genuine working fluency over time.
- Clarify the oversight framework. Decide where AI governance formally sits — audit, risk, or a dedicated technology sub-committee — and document it. Ambiguity about ownership is itself a
governance risk.
Boards that take a structured approach to this will be better placed for the decisions ahead — and for the scrutiny, from regulators and investors alike, that is already on its way.