AI might run companies soon. Corporate boards beware

Douglas Sarro is an assistant professor at the University of Ottawa Faculty of Law.

Ed Waitzer is a retired lawyer is a senior fellow at the C.D. Howe Institute and former chair of the Ontario Securities Commission.

Even the most elaborate corporate values statements and compliance policies can be trumped by corporate culture – unwritten customs about “how we do things around here.”

An independent investigation into the Wells Fargo cross-selling scandal viewed bank staff’s “sales culture” as a decisive factor in their opening millions of accounts without client authorization. In 2017, Volkswagen’s then-CEO cited a culture of excessive deference to authority as a factor behind its staff designing its vehicles to evade U.S. emissions standards. Meanwhile, whistleblowers at Boeing have said the crashes of two of that company’s 737 Max planes can be traced to a culture of cost – and corner -cutting.

Though boards are responsible for overseeing corporate culture, in trying to execute this role they may find themselves in much the same predicament as the courtiers in The Emperor’s New Clothes – recognizing obvious flaws but unable or insufficiently confident to articulate and address them.

Corporate culture is notoriously difficult to define and measure. This challenge is compounded for part-time, independent directors given their capacity and time constraints and their tendency to focus and rely on materials prepared by management and geared to the cadence of a more or less quarterly meeting cycle.

This problem seems only likely to get worse, as AI accelerates the speed of more and more spheres of corporate decision-making, making culture and compliance problems even more difficult for part-time boards to detect and act on before it’s too late.

Opinion: AI robots tempt the caregivers of loved ones with dementia to step aside

Our board oversight model is no longer fit for purpose. To use scientific parlance, boards are increasingly sampling data that is below the rate of change.

Directors should still put a premium on collegiality and the credo of “noses in, fingers out” (i.e., staying informed without micromanaging). But they will need to design protocols for gathering and making use of better-quality information that is subject to independent preparation, verification and analysis.

The surest way for boards to meet this challenge is by leveraging AI themselves. For example, earlier this year, Lloyd’s Banking Group started using a “board bot” to help board members prepare for meetings and check for bias in decision-making.

We expect it will not be long until boards are using AI to gain broader and deeper visibility into corporate activities at lower cost and in real time. Imagine an audit or risk committee where an AI agent continuously monitors financial transactions or compliance metrics – elevating analytics and predictive insights and enhancing the ability of directors to provide effective oversight, strategic judgement and a moral compass.

Directors will not need to become AI experts, but they should be comfortable evaluating processes that are designed to surface AI risks early and that define their own responsible use of AI in discharging their duties. Boards should also be reconsidering the cadence of board processes (and related issues such as board minutes and the maintenance of privilege) in light of the availability of information, analysis and verification in real time.

By enabling boards to manage their time more productively, face difficult discussions more effectively, and navigate risk with more agility, AI should narrow the mismatch between governance ideals and practice.

Many are experimenting with going further, by using AI to replace board members altogether. This summer, Delaware’s Secretary of State proposed testing new “artificial intelligence companies” – to be managed by an AI agent rather than human directors and officers – in a regulatory sandbox.

Coyne: The question isn’t whether we should slow down AI development. It’s whether we still can

Skepticism about AI’s potential to fully replace the judgment exercised by a human board of directors seems warranted. AI does not set out to tell right from wrong or fact from fiction. Rather, as the Archbishop of Canterbury recently observed, it merely “produces a statistical echo of what has been said before.”

But even if AI never fully replaces human boards, we expect it will absorb many responsibilities now exercised by management. In this context, the focus of managerial accountability will likely shift from assigning blame to specific individuals to ensuring there are feedback and response mechanisms in place that define, facilitate and require the rapid identification of problems and their remediation.

As C.S. Lewis put it: “We all want progress. But progress means getting nearer to the place where you want to be. And if you have taken a wrong turning then to go forward does not get you any nearer.” It will fall on boards, as well as legislators and regulators, to be accountable for the path forward.