The Boardroom AI Divide: Agreement on Paper, Friction in Practice
Why chief executives and their directors keep clashing over the speed, ownership, and payoff of artificial intelligence
On the surface, the people running large companies and the people overseeing them seem to be reading from the same script when it comes to artificial intelligence. Both groups insist the technology is a priority. Both say it needs disciplined governance. Both expect it to reshape how their organizations compete. Yet a closer look at how each side actually thinks reveals something more uncomfortable: the consensus is mostly rhetorical. Once the conversation moves from whether AI matters to how it should be run, the cracks start to show.
A recent global study of 625 senior leaders — 351 chief executives and 274 directors at companies generating at least $100 million in annual revenue — mapped exactly where those cracks appear. Just under half the respondents were based in the United States, and roughly three-quarters represented privately held firms. The picture that emerges is one of two camps that broadly trust each other’s intentions while quietly doubting each other’s judgment. Five disagreements stand out, and each one carries the potential to slow an AI effort or sour the relationship at the top of the company.
1. Executives suspect the hype has gotten into the boardroom
When chief executives were asked what they most wished their boards would change, the leading answer was striking: directors need to separate the marketing noise around AI from what the technology can realistically do today. More than half of CEOs pointed to that gap between headline excitement and operational reality as the thing slowing them down. Directors, for their part, returned the criticism in a different form — many said the real problem was that CEOs had not done enough to explain their AI strategy clearly. A recurring theme in their written comments was a desire for executives to communicate their vision more openly and more often.
In other words, both sides feel under-informed by the other. CEOs think directors are reacting to a distorted version of the technology, and directors think CEOs are keeping them in the dark about the plan. Neither perception is flattering, and together they describe a communication breakdown rather than a genuine difference of opinion about AI’s importance.
2. Boards rate their own AI fluency highly. CEOs aren't so sure.
Confidence is not in short supply around the board table. Three out of four directors rated their own understanding of AI as equal to or better than that of their peers. Chief executives see it differently. Roughly a third believe their boards overestimate which human tasks the technology can actually take over, and close to four in ten say directors lack a well-grounded view of how AI is changing the way companies grow. Pulled together, the data suggests a majority of CEOs quietly question whether their boards grasp the technology as well as the boards believe they do.
This is a delicate gap, because self-assessment shapes behavior. A director who feels fluent will push harder, ask different questions, and weigh risks differently than one who feels out of their depth — and if that confidence is not matched by real understanding, the guidance flowing to management can be confidently wrong.
3. Boards want to move faster. CEOs want to move carefully.
Around six in ten chief executives said their boards are pushing AI transformation along too quickly. More revealing is what sits underneath that impatience. Directors who rated their own AI knowledge as weaker were the most likely to feel their organizations were falling behind — a pattern that looks a lot like fear of missing out. Uncertainty, rather than insight, appears to be driving the urgency. The less a director understands the technology, the more anxious they become that the company is not adopting it fast enough.
There is a second twist. The frustration with slow progress runs highest among those who hold the most relaxed views on governance and oversight. That combination — wanting maximum speed while wanting minimum guardrails — is precisely the posture most likely to put a board at odds with a CEO who is trying to balance ambition against risk. Speed and control are being pulled in opposite directions by the same people.
4. Everyone agrees the executive team should own AI. The CEO ends up carrying it.
Ask both groups where AI strategy should live, and they agree: with the executive leadership team as a whole. Fewer than one in ten on either side think a dedicated chief AI officer should be in charge, which reflects a shared belief that the work belongs across the C-suite — the marketing chief, the people chief, the transformation chief, and the rest all have a part to play.
Practice tells a different story. Nearly half of chief executives say they are personally leading AI implementation, a reality that directors acknowledge but understate. The principle is shared ownership; the lived experience is the CEO doing the heavy lifting alone. That mismatch matters, because a strategy nominally owned by the whole leadership team but actually driven by one person is fragile — it bottlenecks at the top and stalls whenever that one leader is stretched thin.
5. CEOs feel judged on AI more than their boards admit
The two camps even disagree about how much AI weighs on the CEO’s own report card. Directors estimate that a little over a quarter of a chief executive’s performance review hinges on hitting AI return-on-investment targets. The chief executives themselves put the figure above a third. That difference between what leaders feel they are accountable for and what their overseers formally measure is more than a rounding error — it means CEOs may be steering toward goals their boards never explicitly set, while boards may be underestimating the pressure their leaders are absorbing.
AI literacy is becoming the price of a board seat
Given how many of these disagreements trace back to uneven understanding, it is no surprise where the two groups finally converge. Roughly eight in ten chief executives and eight in ten directors now believe that anyone joining a board should be able to demonstrate, in measurable terms, that they understand how AI will reshape their industry. What was recently a nice-to-have credential is quickly becoming a baseline requirement. The bar for boardroom membership is being redrawn around fluency in the technology, not just experience with the business.
Closing the gap
None of these five fault lines is fatal on its own. What makes them worth watching is that they all point in the same direction: the relationship at the very top of the company is running on assumed alignment that does not survive contact with detail. The fix is not more enthusiasm for AI — there is plenty of that already — but a shared, honest baseline of understanding, clearer communication about strategy, explicit agreement on who owns what, and accountability that both sides have actually signed up to.
Organizations that close these gaps early will spend their energy on the technology itself. Those that ignore them will spend it managing tension in the boardroom — and at a moment when coordinated leadership is exactly what an AI transformation demands, that is a cost few companies can afford.
