A Goldman Sachs executive breaks down how AI can be used in corporate boardrooms

As the generative AI revolution unfolds, we’ve witnessed the emergence of AI copilots, AI teammates, and even part-time AI CEOs. 

This increasing permeation of AI into the fabric of our business institutions begs the question as to when we will see AI establish a presence in the corporate boardroom. While it is hard to imagine a world in which AI replaces human board members in total, it seems clear that AI-enabled tools will impact the future of corporate boardrooms. 

I’m the cohead of the Goldman Sachs Global Institute, and I believe AI can help cultivate a culture of accountability, candor, and effective governance. Here’s why:

AI as an impartial spectator 

My colleague Marco Argenti recently made the case that, in the age of AI, more students need to study philosophy. One work they should pick up is Adam Smith’s less-known book, The Theory of Moral Sentiments. Smith wrote of the concept of the “impartial spectator,” a hypothetical, well-informed observer whom he describes as “reason, principle, conscience . . . the man within, the great judge and arbiter of our conduct.” It’s a trusted, objective observer, a devil’s advocate, a voice of reason. AI may be uniquely capable of fulfilling this dispassionately logical role, including in corporate boardrooms.

Consider the responsibilities of corporate board members—broad oversight, ensuring proper governance, and loyalty. These obligations demand the synthesis of vast amounts of information, the ability to weigh conflicting evidence, ask challenging questions, and render guidance in complex environments. Today’s AI-enabled systems are increasingly capable of fulfilling these roles and adding their “voice” to board deliberations.

I realize the thought of introducing AI-enabled systems into a corporate boardroom might seem discordant, as boards are composed of experienced individuals with specific, relevant expertise—the sort of people you might think of as irreplaceable by machines. It may be better to think of AI not as replacing these capabilities, but as a tool to augment them, empowering boards to perform their responsibilities with the assistance of a tireless, keen, and unemotional partner.

This partnership can take many forms. Imagine applying future AI-enabled systems toward some of the most difficult challenges faced by boards today, first by diversifying the insights any board can have access to. This could entail an AI-enabled system that is grounded in the salient facts surrounding any given board matter or business decision, one that can analyze market trends, customer reviews, and competitive dynamics in minutes. Such a tool could broaden the context of any discussion, introduce novel insights, and connect the dots across complex scenario analyses—and do so free from the human constraints of limited recall. This capability can provide board members with data-driven insights that would help in making more objective decisions grounded in evidence, rather than intuition or partial perspectives.

Raising new questions

AI-enabled systems could also expand the perimeter of boards’ awareness, raising questions they didn’t know they needed to ask. The breadth and range of a well-tuned AI could enrich board discussions by posing a wide array of questions that help to enhance oversight and pressure-test decisions. While human board members bring a range of experiences and insights, the complexities of today’s environment demand expertise that is both broad and deep. An AI could integrate a wide range of knowledge and pose important questions that help to identify new risks and expand the envelope of issues to consider.

There are also natural human tendencies, such as interpersonal power dynamics, that sometimes prevents new insights from surfacing or new questions from being asked. For boardrooms, these power dynamics can stifle debate, as the nature of some uncomfortable human relationships often challenges board members in carrying out their duties of candor. Personal sensitivities, subtle exercises of power, and varying degrees of tenure and expertise can inhibit direct and substantive debate. An AI, unburdened by these natural influences, could help ensure that “unspoken concerns” are voiced.

Planning for the unexpected 

AI-enabled tools could aid boards as they plan for the unexpected. When he was asked what the most difficult part of being prime minister was, Harold Macmillan reportedly responded, “Events, my dear boy, events.” This decade, corporations have been confronted with a series of events that were once difficult to imagine. For instance, a pandemic and the rise of AI. To assist them as they prepare for an uncertain future, boards can consider using AI-powered scenario planning tools to help them navigate the complexities of global operations. An AI-enabled system can simulate the impact of political changes, supply chain disruptions, or technological innovations on company operations. By asking critical “if this, then what” questions, or “what risks have we not yet considered,” AI could help boards in making dynamic risk assessments and creating alternative strategies.

Room for improvement

Of course, today’s AIs are far from perfect, and there should be a wide variety of considerations before introducing an AI-enabled system to the boardroom. There are limitations to this technology, including hallucinations, a lack of explainability, and the tendency to mimic other human biases expressed in training data. Such factors should cause board members to consider AIs as boardroom companions, rather than as voting members or independent thinkers, and to process AI inputs and outputs with a healthy understanding of their current limits and the need for continued improvements. It’s incumbent on boards to ensure that they evaluate and implement appropriate guardrails and mechanisms so that they can safely leverage AI capabilities.

These imperfections do not change the fact that part of a board’s duty is to harness innovation to the benefit of the company and its shareholders. Another critical responsibility of boards is to recognize the context in which their companies operate and adapt to changes. In corporate America, such adaptability is needed now more than ever, as boards deal with today’s complex and changing business environment, navigating evolving economic, geopolitical, social, and technological realities.

It is the mission of boards to chart a way forward for their companies. AI-enabled tools may provide exactly the kind of impartial spectator and new perspectives that Smith wrote about more than 200 years ago. Humans and machines, working together, could turn a philosophical concept into a reality.