AI Might Run the Business, but You Can't Automate a Board

There's an old joke about the factory of the future: it will be run by a human and a dog. The human's job is to feed the dog. The dog's job is to keep the human from touching the machines. That whimsical image takes on new weight in light of OpenAI CEO Sam Altman's remark that he anticipates a "10-person, $20 billion company soon."

AI makes it possible to do more with fewer labor hours. Execution becomes faster and more consistent. Product cycles accelerate. Costs decline. For many executives, these improvements signal progress. But there is a limit to what can be streamlined. An operation might survive with a skeleton crew. Governance cannot. It depends on experience, expertise and judgment.

Experience recognizes patterns, learns from the past, and spots outliers and averages. Expertise brings deep domain understanding. Judgment weighs incomplete information, ethical trade-offs and long-term consequences. Together, these three forces allow boards to lead where technology cannot. Companies can automate code. They cannot automate the corporate conscience.

In times of change, humans often overinvest in speed and underinvest in discernment. As executives integrate AI operationally, boards must evolve their oversight. Good governance in the age of AI requires both technological progress and human governance to move together: challenging, informing and reinforcing one another.

Boards Need to See AI as a Shared Leadership Issue

We are often asked to advise boards as designated "AI experts." Our advice is always the same: AI cannot be confined to a single seat at the table. It must be a shared responsibility of everyone present. While these invitations reflect growing awareness, they also reveal a deeper misconception: that AI is primarily technical rather than strategic.

At a recent gathering of directors and CEOs, we observed a familiar pattern. Leaders in attendance had skimmed reports and attended AI briefings. Still, there was unease. Staying informed isn't the same as being equipped. Preparing boards for AI governance takes more than updates. It requires rethinking how CEOs and directors engage on questions of judgment, risk and long-term value.

Conversations about AI often begin in IT and rarely reach the boardroom with the clarity they deserve. Some focus on compliance, others on product demos. What is often missing is recognition that AI is a matter of organizational direction and leadership. Helping boards understand how AI touches product, talent, regulation, capital and ethics isn't about technical mastery.

Boards that govern AI well are those that foster challenge, curiosity and reflection. In machine learning environments, where outcomes are opaque and probability-driven, directors must probe assumptions, debate implications and navigate uncertainty. That requires a culture where disagreement is expected and inquiry leads to accountability.

Too often, AI is treated as a technology investment or operational risk. Responsibility gets assigned to one executive or offloaded to a committee. But governance isn't passive. And in a world defined by automation, doing nothing is still a decision.

Give Boards a Wider Lens

In our experience, boards aren't indifferent to the challenges faced by AI. They're simply operating with governance models designed for a different era. Traditional governance assumes risks can be mapped, data flows are contained, and accountability is linear. AI upends all of that. It generates probabilistic outcomes, emergent behaviors and blurry boundaries between partners and platforms.

To lead effectively in an AI-driven enterprise, CEOs must help boards modernize oversight. That means investing in fluency, not just reporting. It means asking better questions: Are our leadership pipelines adapting? Is our culture resilient enough for distributed decision-making? Are employees empowered to escalate when values are at risk?

Most boards are not yet ready. A 2024 BCG report found that few boards have frameworks to assess AI's ethical, reputational or strategic risks. CEOs and boards must bridge that gap. Frame AI as a systemic force, not a stand-alone tool. Embed it into risk, strategy and culture discussions. Make AI oversight a regular dialogue, not a one-time presentation. Help boards build the muscle to ask where technology is making decisions — and what happens when those decisions go wrong.

Put Culture and Judgment at the Center of Decision-Making

AI doesn't just change workflows; it reshapes how people perceive responsibility, authority and purpose. Culture becomes the connective tissue between what technology enables and how people behave.

As AI systems make ever more recommendations and decisions, the potential for drift, whether in mission, in trust, or in values, grows. CEOs must work with boards to ask: How are we maintaining ethical clarity as decisions disperse? What values are embedded in our systems? Who speaks up when something feels off?

Regulators will define the legal boundaries. But the most consequential decisions will happen inside organizations. Boards and CEOs must combine legal guidance with moral clarity and institutional leadership.

Building Strong Boards for the AI Era

Some boards are evolving. They are creating Responsible AI committees, running scenario planning, and adding new members with technical backgrounds. These are promising signs. But modest initiatives are not enough. The best boards in the age of AI won't be those that adopt technology fastest. They'll be the ones that cultivate experience, foster expertise and practice sound judgment. These are not efficiencies. They are investments. And they are the foundation of governance that endures.