The sweet taste of a new idea
Just as a perfect cookie offers a delightful sensory experience, so too does a truly novel idea for behavioral economist Sendhil Mullainathan. The Peter de Florez Professor at MIT, with joint appointments in Economics and Electrical Engineering and Computer Science, and a principal investigator at the MIT Laboratory for Information and Decision Systems (LIDS), Mullainathan describes encountering new insights with the same “hedonic pleasure” one gets from a breakthrough after being stuck on a problem.
Mullainathan’s unique cognitive approach, which he describes as “out of phase” with conventional thinking, began early in life. As a child, he recalls seeing multiple valid explanations for every multiple-choice answer on tests, a perspective that set him apart. This unconventional mindset, rather than being a hindrance, has proven invaluable in his research, earning him accolades such as a MacArthur “Genius Grant,” designation as a “Young Global Leader” by the World Economic Forum, and recognition as a “Top 100 thinker” by Foreign Policy magazine, among others.
A pivotal influence on Mullainathan’s career was a childhood experience with financial precarity. At age 10, when his father lost his job, he grasped the fragility of economic security, an understanding that profoundly shaped his focus on financial scarcity. This personal insight drew him away from standard economic models and towards behavioral economics, a field that incorporates the complex, often irrational, aspects of human psychology into economic decision-making. He was inspired by early pioneer Richard Thaler, later a Nobel laureate, noting that the “non-math part of this field that’s fascinating” is precisely where traditional economic math falters because “people are weird and complicated and interesting.”
His commitment to understanding human behavior led to groundbreaking empirical research. In 2013, Mullainathan co-authored the influential paper “Poverty Impedes Cognitive Function,” published in Science. The study, involving sugarcane farmers, demonstrated a significant decline in cognitive performance when individuals faced severe financial hardship before harvest, compared to their improved scores post-harvest when financially stable. This research had a far-reaching impact, increasing policymakers’ awareness of how economic precarity taxes cognitive resources, making program sign-ups or complex decisions difficult. More importantly, it resonated deeply with individuals who had experienced economic insecurity, validating their struggles and offering a scientific explanation for what they felt.
Mullainathan’s recent return to MIT, where he previously taught, signals his deep dive into artificial intelligence and machine learning. His goal is to bridge computer science with behavioral economics, seeking to expand human capabilities rather than merely automate existing tasks. He challenges the field to move beyond simple automation, urging researchers to ask: “What capacity do you want expanded? How could we build an algorithm to help you expand that capacity?” He envisions AI as a powerful tool for augmentation, pushing the frontiers of fields like psychology and economics.
He fundamentally believes that “the next generation of breakthroughs is going to come from the intersection of understanding of people and understanding of algorithms.” One compelling application he describes involves AI providing decision-makers—such as judges or doctors—with insights into their “average” decision for a given scenario. This “average-you” metric could help individuals assess if their current decision is being influenced by transient factors like mood or stress, encouraging a more reasoned approach. Mullainathan summarizes this powerful concept: “average-you is better than you.”
For Sendhil Mullainathan, the pursuit of such innovative ideas is not just academic; it’s driven by the profound and “delicious reward” that comes from discovering new ways to understand and improve the human condition, amplified by the transformative potential of AI.
