Boards have massive AI governance deficits which they can’t afford

Boards have massive AI governance deficits and, at the rate the technology is intruding on corporate life, they can’t afford to do so for much longer. 

To most people in the governance space, it won’t even come as a surprise. But that shouldn’t take away from stark statistics showing how much work still needs to be done in bringing structure to a chaotic time. 

Without it, we rapidly end up entertaining levels of risk that directors are supposed to manage with confidence. Sure, AI is a different challenge when it comes to risk – and perhaps the biggest many directors will ever face – but the drive for control shouldn’t be lost. 

Sadly, the statistics point to just that right now:

These figures give you an idea of the state of play

1. According to a 2026 PwC survey, 71% of directors say their boards need to strengthen their AI skills. 
2. According to Deloitte, 66% of directors continue to have “limited to no knowledge or experience” with AI. 
3. According to OnBoard, 92% of board directors used AI for board work in the last six months, despite 60% of them sitting on boards that lack any AI policy. 
4. Our own research at The Corporate Governance Institute showed that, for most boards (between 80-95%, depending on the sector), implementing new AI tech is a short-term priority (within 12 months).

So, in summary, there’s incredible appetite for change but a very worrying lack of any structure to make that appetite safe.

It’s important to note that most of these statistics are coming from boards themselves. It’s not external assessment, but internal reflection – and a lot of it lays the lack of structure bare for stakeholders to see. 

Zoom in, and the picture gets clearer on a fundamental disconnect. On one side is AI, its widespread use, its broad set of opportunities, and its evolving set of associated risks. On the other side are boards, entrusted with even more fiduciary responsibility in the modern era, and subject to intense scrutiny. 

Those two sides simply don’t link up as much as they should. 

Metrics are a clear indicator. According to PwC’s statistics, many boards are reporting poor, or even non-existent, metrics when it comes to linking AI outcomes, risks, and business performance (82%), return on investment in AI (75%), and workforce readiness to embed AI across the organisation (69%). 

It all adds up to boards being very embracing of AI, but lacking any grasp of how it will achieve what they think it will achieve. There are no milestones, no broken-down targets. Everything is high ambition backed by low understanding. 

The alarm-bell moment is that this is “Corporate Governance 101”. Directors are supposed to make decisions based on clear reasoning and oversee their implementation in the form of strategic goals. When those elements are absent, it’s a problem.

AI is different. Commentators, some of whom we have interviewed exclusively at The Corporate Governance Institute, have singled out the remarkable pace that AI moves, and how traditional corporate structures just weren’t built to match it. 

Technologies change, opportunities evolve, risk becomes harder to manage. We’re seeing far clearer examples of it with AI than we have done with other big innovations that have swept businesses into new eras. The internet, the drive for sustainability, financial crises – none of them has forced a change in outlook as fast as AI is doing. 

So, it’s important to give boards their due credit in that context. In fact, a board that is open about its need to improve its AI credentials should be commended for acknowledging reality to itself and to key stakeholders. 

The trick, now, is to take something from it.

AI is a wave of change that dramatically outpaces even the fastest governance structures. It’s not just directors; it’s lawmakers and regulators as well. There’s just too much, too new, at once. 

As you might expect, there’s no simple fix that will futureproof your company for as long as AI continues to evolve at this pace. Instead, it’s about preparedness wherever and whenever possible. Adaptation is a key strength; can you and your business evolve in six months when some new tech becomes widely available, or new rules come into force?

For this, it is obviously desirable to have people on your board who know AI and respond to its evolving potential with practical input into decision-making. However, this is admittedly not always possible; AI is still a new phenomenon (although it might not feel that way a lot of the time). Instead, one of your primary avenues to AI literacy should be upskilling your current directors with a baseline level of knowledge. 

You can afford not to know everything about AI’s potential over the next ten years. You cannot afford to oversee implementation of AI and remain in the dark on goals or how you’ll measure them. That’s the kind of thing that baseline literacy will help address. 

AI should never be a blind jump, and the companies that move away from that way of thinking first will be better placed to show every single stakeholder that they have an acceptable level of control.