Every well-run board reviews its committee work at some point in the year. At one board's annual effectiveness review, that routine exercise threw up a question nobody had a ready answer for: where does AI governance actually fit into the work of this board?
Is AI governance full board business? Committee business? Some deliberate mix of both?
The Chair put the question to the room. Everyone had an instinct. The Corporate Governance Chair and the Senior Independent Director had not fully considered the matter. Nobody could point to a moment when the board had actually debated this.
That's the real gap. Not who owns a particular AI risk. Who owns AI governance itself, as a standing part of how this board is governing the company.
Efficient is not the same as competitive
Before we get to where AI governance should sit, it's worth asking what it's actually there to test.
Edition 9 gave you a real example. A CEO said AI made the company the lowest cost operator in the market. That sounded like strategy.
It wasn't. It was efficiency, dressed up as strategy.
A lower-cost version of the same old business can still lose. A rival with a better offer beats you on value, not on price. Savings buy time. They don't buy position.
The governance test is: does AI improve our competitive position for the long term?
Nobody decided. It just happened.
AI governance is not being designed in the boardroom. Selective briefings by C-Suite leaders on aspects of AI have shaped board activity related to AI. Challenges with directors' and chairs' capacity to understand AI have compounded the issue.
The key challenge is that AI's features have become a distraction for talented governance leaders. It's time to reduce the distractions and get on with governing this high-capacity resource.
Governance-grade questions need to be shaping discussions. Is AI enabling our strategy, and improving our competitive position? And, if so, how do we optimise it and manage it like a valuable resource?
Part of a longer arc
In this six-part series on AI Governance, this edition builds on the last three (Governance Decoded Editions 7 to 9), rather than starting fresh.
Edition 7 gave you a working grasp of what AI actually is, so the AI governance conversation could move past buzzwords. Edition 8 turned that understanding into five domains of sharp, board-level questions. Edition 9 took one of those questions further still, into a live strategic risk: a dependency on a single AI provider for too much of the business.
This edition is the next step in that same progression. Once you can ask the right questions and spot the risk, where does the responsibility for holding that AI governance conversation actually sit on your board?
Three places AI governance can live, and what each one misses
A standalone AI committee builds depth. The directors on it learn the technology well. They are normally briefed by the Chief Information Officer, not the CEO.
Unlike Audit and Risk, a committee like this doesn't usually appear by accident. Someone decides to create it. But look closely at the decision behind it. It's rarely "let's govern our competitive position." It's usually "let's get smarter about the technology, fast." A committee can be built on purpose and still be built to answer the wrong question. Ask "does this work" enough times, and you stop asking "is this the right bet."
Audit oversees controls and financial reporting. It tends to look back, not forward, with questions about whether you did what you said you'd do. It rarely asks whether you should have done it at all, given where the market is heading. The dependency risk from Edition 9 isn't really a controls question. It's a strategy question.
Risk oversees downside more than upside. It asks what could go wrong. It doesn't naturally ask what could be won. AI brings opportunity as well as risk. A committee with a risk management mindset is unlikely to be the best fit for holistically assessing opportunities.
I've sat on all three types of committee. None of them will ask whether the AI in use is actually good for your competitive position.
Add them up. Is that enough?
Some boards try to fix this by spreading AI governance across all three. Give the AI committee the depth. Give Audit the controls. Give Risk the downside. Does adding their work together actually give you strategic governance of AI?
AI touches strategy, operations, the workforce, and the cost base, all at once. Add three partial views together and you still don't get one whole view.
Splitting AI governance across committees isn't wrong. It may just not be enough. The answer is to govern AI as a strategic resource, not as something that some committees can partially address.
This is not a technology discussion. It's a leadership one.
The governance of AI cannot be left unaddressed while AI use keeps expanding inside the business. Surveys this year consistently reveal that about half of boards haven't yet had a proper debate about how AI is governed at board level. For most companies that haven't had that debate, it's now overdue.
That work sits with the Chair because the Chair is the one person who can decide to lead the board on this strategic matter.
Leading this isn't a technology skill. It's closer to the skill a Chair already uses when the board debates hiring a senior leader into a role the company has never needed before.
If a Chair can lead that discussion, they can lead this one. The skill required is the same skill needed for strategy, for competitiveness, for any future-facing question that doesn't sit neatly inside one committee's brief.
Prompt, before it drifts
You don't have to wait for the Chair to get there alone.
Raise it directly: has the Chair got a view on whether AI governance is in place, wherever the pieces currently sit? If it's split across committees, has anyone actually checked whether their combined work adds up to a strategic view, or has everyone assumed someone else has it covered?
Then go a step further. Offer to help. A Chair who hears "here's how I could help you think this through" moves faster than a Chair who only hears a warning. You don't need to be an AI expert - just a shaper of good governance.
Language / Clarity / Confidence.
This edition exists so you can walk into your next board meeting with the language, the clarity and the confidence to be able to see a clear path to governing AI.
Language. Carry the language that focuses on AI Governance, not just AI.
Clarity. Seek clarity as to whether your Chair is leading how AI governance works at board level. Is it being shaped by board members, or allowed to happen in whatever way the C-Suite chooses to brief the board?
Confidence. You need the confidence to raise this with your Chair, offer to help carry it, and note the answer even when the answer is an uncomfortable silence. Given the fiduciary duty you carry personally under applicable due care and loyalty standards, raising this isn't overreach. It's good governance.
Boardroom Takeaway
Two things to take into your next board meeting.
For the Chair: "How we govern AI and its fit with our strategy. I want to decide where AI governance sits, with no coverage gaps."
For you as a board member: offer to help the Chair get there, not just raise the flag.
This is a test of board leadership. If your board's leadership isn't yet equipped to lead a discussion like this one, on strategy, on competitiveness, on where the company is actually headed, that's the more urgent repair. Fix that first.
If you're the Company Secretary or General Counsel reading this: you're often best placed to help the Chair get this onto the agenda formally.
Until next time,
Paul
This edition maps the board's role in an AI incident — not the management response, but the governance layer. Part 5 of 6.