Photo by Tima Miroshnichenko from Pexels

In a Harvard Business Review article, the institution digs deep into a recent experiment relating to an intriguing question: Can artificial intelligence (AI) replace a CEO? The findings make it clear that leadership requires more than just number-crunching—it involves foresight, intuition, and the ability to manage disruptions, qualities that AI currently lacks and may never possess.

Researchers Hamza Mudassir, Kamal Munir, Shaz Ansari, and Amal Zahra, led the experiment, which ran from February to July 2024, to test whether AI could outperform human CEOs in a simulated environment. The participants were composed of executives from South Asian banks, business students from Central and South Asian universities, and OpenAI‘s GPT-4o.

The task was to strategize decisions for a digital twin of the U.S. automotive industry; the end goal was to maximize profitability and survive as long as possible without being fired by a virtual board.

The decision to run this experiment reflects a growing interest in AI’s role beyond mere data processing and automation. Companies today are increasingly integrating AI into decision-making processes.

That’s the same AI that has been telling businesses to break the law or hallucinate events and critical information. Thus, the findings bring up a more pressing question: Why test AI in roles that require human intuition, especially when its limitations are so apparent?

AI’s Short-Term Wins, Long-Term Challenges

The AI model’s inability to think beyond the immediate future should be a red flag for any company considering AI for leadership.

In the simulation, GPT-4o succeeded early on by optimizing prices and products with data-driven precision, but once the market shifted, its rigid approach led to failure. Human participants, particularly the top-performing students, took a cautious, long-term approach to strategy, which allowed them to navigate unpredictable disruptions.

GPT-4o, on the other hand, was fired from its virtual CEO role more quickly than the humans.

This underscores why companies may not be able to rely solely on AI to lead them through crises. While AI has excelled in certain areas, such as operational efficiency and market responsiveness, it remains uncertain whether it can navigate unpredictable “black swan” events that require human intuition.

Another major concern is accountability. AI can make data-driven decisions quickly, but when those decisions lead to failures, who takes responsibility?

The Future of Leadership: Hybrid Models

The findings suggest that AI is more suited to augmenting human leadership rather than replacing it entirely. AI can process massive datasets and optimize short-term gains, but it lacks the human capacity for judgment, empathy, and ethical decision-making—qualities essential for a CEO.

The Harvard article further states, “The future of leadership is hybrid, where AI complements human CEOs, focusing on vision, values, and long-term sustainability. ” This hybrid model refers to a situation where AI supports human decision-makers. It could practically be the future of corporate leadership—a model that allows companies to benefit from both AI’s data-driven insights and human intuition.

Harvard’s experiment offers a window into AI’s capabilities, but it also highlights its weaknesses. It is clear that AI has a role to play in the C-suite, but just not as a full replacement for human leadership. AI may excel in specific areas of corporate strategy, but its limitations in handling uncertainty and ethical dilemmas mean that human judgment will remain irreplaceable, at least for now.