The boards that close this gap fastest will be those that have developed the collective intelligence to govern in conditions of genuine uncertainty.
There is a gap opening up between what AI is doing inside organisations and what boards are equipped to govern. The technology has moved into operations, into customer-facing decisions, into strategic forecasting and into the daily workflow of staff. Board oversight has not moved with it.
The data describes the shape of the gap. More than 88 percent of organisations are actively using artificial intelligence in some form. Only 39 percent of Fortune 100 boards disclose any structured oversight of it. Two thirds of directors globally report limited to no knowledge of the technology. One in three says AI does not reach the board agenda at all.
Two recent reports have brought the picture into sharp focus. A Deloitte memorandum published through the Harvard Law School Forum on Corporate Governance (Marks, Abrash and Probst, 2025) and a McKinsey report on board readiness for AI (Baig, Dave, Huber and Vuppala, 2025) have, independently, arrived at the same uncomfortable conclusion. Boards are governing organisations that have already been transformed by AI, and the board may not have been able to catch up yet.
For directors personally, this matters. Fiduciary duties do not pause while a board develops its AI literacy. Regulators, shareholders and litigants will, in due course, ask what the board knew, when it knew it, and what it did about the gap. The Deloitte data is particularly telling here. Seventy-two percent of boards engage their Chief Information Officer on AI matters. Twelve percent engage their Chief Risk Officer. That asymmetry will be a difficult position to defend in hindsight.
The numbers describe a governance maturity problem.
A Maturity Question
The instinct, when confronted with this data, is to treat it as a knowledge deficit. The proposed solutions follow predictably. More training. More expert directors on the skills matrix. More external briefings. More frequent updates to the board.
These are necessary. They are not sufficient.
There is a four-stage Ascension Ladder of governance maturity that helps a chair locate where a board currently sits and identify the development pathway from where it is to where it needs to be. The model draws on published coaching research on practical wisdom (King, Norbury and Rooney, 2022).
Stage 1: Reactive Governance. The board responds to events as they arrive. There is limited shared sense-making between meetings, and the board's attention is shaped largely by the papers in front of it.
Stage 2: Personal Awareness. Individual directors are doing their own homework on emerging issues. They read widely, attend external briefings, and bring their findings into the room. But there is no coordinated learning, no shared framework, and no collective integration of what each director knows.
Stage 3: Cognitive Alignment. The board has developed shared language, common mental models and integrated risk thinking. Directors can hold a conversation about complex emerging issues without first having to define terms or contest premises.
Stage 4: Collective Wisdom. The board operates as an integrated intelligence. It senses weak signals, holds competing truths, and makes decisions that reflect the full complexity of the moment, including the dimensions that have not been formally tabled.
What the two reports describe is a population of boards stalled between Stage 1 and Stage 2. The 59 percent of directors who have sought AI education on their own initiative, without coordinated board-level training, are a textbook portrait of Stage 2. Everyone is doing their own homework. No one is doing it together.
This matters because AI governance is a team sport. The risks it presents, including ethical ambiguity, algorithmic bias, vendor dependency and strategic misdirection, are systemic. They require Group Attunement, the board's collective capacity to sense unspoken dynamics, read the emotional climate of the room, and act as an integrated intelligence.
The 72/12 Pattern Is Groupthink in a New Costume
The 72 versus 12 finding from the Deloitte report deserves more attention than it has received.
It is the most predictable source of future governance exposure on AI matters. An overreliance on technology officers, with risk officers consulted only intermittently, will be a difficult position for boards to defend in hindsight to regulators, shareholders, or litigants.
It is also a recognisable pattern in the governance work I have observed. The asymmetry is produced by the social pressures that quietly accumulate in any high-performing group. This is groupthink in a new costume. Boards gravitate toward the voices that affirm innovation momentum. The voices that complicate it, including risk, security and ethics, become structurally marginalised.
The pattern compounds. When the Chief Risk Officer is consulted only on isolated matters, the board loses the running risk dialogue that should accompany every major AI decision. When the Chief Information Security Officer is brought in late, the board sees security as a brake on innovation. When ethics is treated as a separate workstream, it becomes detached from the commercial logic that drives the decisions ethics is meant to shape.
The antidote is perspective-taking paired with the search for integration. These are the capacities that allow a board to hold competing truths and reach decisions that have been genuinely tested.
This connects directly to intellectual humility, the recognition of the limits of one's own knowledge. In a boardroom saturated with AI optimism and competitive anxiety, intellectual humility is what allows a director to say, plainly and without apology: I do not fully understand this, and that is material information for this board.
The Quality of Attention in The Room
Most of the governance prescriptions on offer focus on structure. Clearer committee mandates. Updated board composition. AI fluency training. Metric dashboards.
Structure is necessary. The quality of attention directors bring into the room is what makes structure work.
Recent research with Justin Brienza and Vincent Murdoch validated the Mindful Board Assessment Survey, an instrument designed to measure two distinct dimensions of attention within the boardroom context (King, Brienza and Murdoch, forthcoming).
The first dimension, Embodied Mindfulness, captures a director's internal present-moment awareness during a meeting. This includes the capacity to notice one's own cognitive shortcuts, regulate one's emotional responses, and maintain focused attention through complex agendas. It is the inner condition for sound judgement.
The second dimension, Group Attunement, captures the board's outward sensitivity to its own collective dynamics. This includes the capacity to read the room, to notice unspoken agendas, to register shifts in tone, and to sense when a decision is being driven by social pressure rather than evidence. It is the condition for genuine collective intelligence.
The analysis confirmed these as two distinct dimensions. A director can score high on one and low on the other. So can a board. Both dimensions behave like skills. Boards that engage in structured development show measurable change.
This is the finding that matters most for chairs. It converts wisdom from an aspiration into a development objective.
From Reactive to Wise
McKinsey offers six governance actions for boards seeking to close the AI oversight gap. All are sensible. All are structural.
A board that installs an AI committee without first developing the collective capacity to sit with ambiguity, challenge its own assumptions and integrate dissenting views will produce a more elaborately governed version of the same problem.
Structure with wisdom becomes effective governance.
Wisdom with structure becomes generative practice.
The boards that close the AI oversight gap fastest will be the boards that have developed the collective intelligence to govern in conditions of genuine uncertainty. To notice what they do not know. To hear from the people they have been inadvertently ignoring. To make decisions that reflect the full complexity of the moment.
That is a wisdom question. The research shows wisdom is measurable. It is also developable. The work begins when a board decides to do it.
Take the Next Step
Where does your board sit on the Ascension Ladder?
If you are a chair or lead director ready to move your board from reactive oversight toward collective wisdom on AI and other complex matters, the Mindful Board Assessment Survey is the place to begin.
Book a confidential conversation → or email elizabeth@drlizking.com.
Related Reading
The Reel Governance series examines collective decision-making and board behaviour through close readings of contemporary film and television.
References
Baig, A., Dave, A., Huber, C., and Vuppala, H. (2025, December 4). The AI reckoning: How boards can evolve. McKinsey & Company. Available at: https://www.mckinsey.com/capabilities/mckinsey-technology/our-insights/the-ai-reckoning-how-boards-can-evolve
King, E., Brienza, J. P., and Murdoch, V. (forthcoming). Measuring mindfulness dimensions in the boardroom.
King, E., Norbury, K., and Rooney, D. (2022). Coaching to develop leadership for healthcare managers: A mixed-methods systematic review protocol. Systematic Reviews.
Marks, A., Abrash, L., and Probst, A. (2025, May 27). Governance of AI: A critical imperative for today's boards. Harvard Law School Forum on Corporate Governance. Available at: https://corpgov.law.harvard.edu/2025/05/27/governance-of-ai-a-critical-imperative-for-todays-boards-2/