Never before in human history have machines entered the sanctum of the boardroom. From Mesopotamian clay tablets that recorded trade to the printing press, the telegraph, and the internet, every invention extended human capability. Artificial intelligence (AI), however, is different. It is not merely another tool of efficiency; it is the first technology capable of generating ideas and making decisions by itself. That makes it more consequential than the telegraph, the printing press, or even the internet. For corporate leaders, AI is not just a new gadget—it is a redefinition of strategy itself.
This raises unsettling questions: If machines can surprise us in games, what happens when they make business decisions with real world consequences?
For CEOs, the dilemma is real: Should they embrace highly accurate but opaque models, or weaker ones that humans can interpret? The trade off between performance and accountability is itself a strategic decision. And unlike traditional business risks, AI’s mistakes can scale at frightening speed.
But AI does not merely execute strategy—it interprets it. That interpretation can go disastrously wrong. Facebook’s engagement algorithms discovered that outrage and hate speech drove more clicks than compassion. Optimising for engagement, the system amplified division and distrust—creating financial, reputational, and social costs. The lesson is clear: Goals given to AI must be framed with extreme care.
First, AI as an assistant. Here, machines handle routine tasks—insurance companies already use AI to process millions of claims. This frees leaders to focus on broader strategy.
Second, AI as master. In this dystopian future, CEOs become overly dependent on AI, delegating critical decisions to opaque systems. Nick Bostrom, in Superintelligence, warned of machines pursuing goals misaligned with human values. The anxiety resurfaced in 2023 when Elon Musk, Yoshua Bengio, and other technologists signed an open letter urging a pause on frontier AI development. The threat is not science fiction: A narcissistic CEO could exploit AI’s real time surveillance power to micromanage employees down to their keystrokes.
Finally, AI as a thought partner. This is the most promising model—where AI augments human judgment without replacing it. But it requires CEOs to act responsibly.
First, benevolence. AI will give leaders unprecedented access to employees’ personal opinions, preferences, and even search histories. With such power comes a responsibility to protect, not exploit, this data.
Second, decentralisation. Information must not be monopolised at the top. Multiple channels of decision making can provide checks and balances against misuse.Third, mutuality. If leaders know more about their employees, employees must also be able to see how decisions are made at the top. Transparency will be vital for trust in the age of algorithmic governance.
The boardroom is no longer just a room of people. It is now an ecosystem where human judgment meets machine intelligence. The ultimate test of leadership in this new era will not be making every decision alone, but knowing how to decide with machines. Those who succeed will not be the CEOs who surrender to AI, nor those who resist it blindly, but those who learn to share power wisely.
Prof. Saptarshi Purkayastha, Professor of Strategic Management at IIM Calcutta
First Published: Mar 04, 2026, 12:02
Subscribe Now