Sendhil Mullainathan: The AI Economist

Source details

Profile of MIT behavioral economist Sendhil Mullainathan by Bob Simison in F&D Magazine, December 2025.
Central thesis: AI is "too important to leave to computer scientists" and economics is uniquely suited to integrating AI with human decision-making.
Mullainathan argues algorithms can dramatically improve human decision-making across domains: job search, bail setting, cardiac test interpretation, and teaching.

Born in a village in the Indian state of Tamil Nadu, south of Chennai.
Family moved to Los Angeles when Mullainathan was 4; father worked in aerospace, later lost job when Mullainathan was 10 due to a security clearance requirement.
Attended the Clarkson School in Potsdam, New York, beginning as a 16-year-old.
Undergraduate studies at Cornell: majors in mathematics, computer science, and economics.
Entered the computer science doctoral program at MIT but deferred to Harvard’s PhD program in economics; completed doctorate in 1998.

Research areas: psychology of memory, decision-making by judges, economics of scarcity, CEO compensation, labor-market discrimination.
Notable works and impacts:
- 2001 finding (with Marianne Bertrand) that "CEO pay responds significantly to luck."
- 2004 paper (with Marianne Bertrand) "Are Emily and Greg More Employable than Lakisha and Jamal?" — finding: White-sounding names received 50 percent more callbacks.
- 2013 book (with Eldar Shafir) Scarcity: The New Science of Having Less and How It Defines Our Lives — core finding: scarcity soaks up cognitive bandwidth, trapping people in scarcity cycles.
Citation metrics reported: body of work cited almost 100,000 times; more than a dozen works with several times 1,000 citations each.
Academic appointments: Harvard, the University of Chicago, and MIT.

Began applying machine learning to human decision-making in 2012; sought topics "well off the beaten path."
2017 study (with four colleagues) on bail-or-jail decisions using machine learning on a database of more than 700,000 people arrested between 2008 and 2013 in New York City:
- Finding: judges routinely made the wrong call; evidence of gambler’s fallacy influencing decisions.
- Estimated impacts: using a risk-assessing algorithm could help reduce crime by 25 percent with no change in the number of people held in jail, or reduce jail populations by 42 percent with no increase in crime.
- Outcome: researchers built an AI tool that New York City judges use today to aid decision-making.
2024 paper (with Jens Ludwig) using AI on defendant mugshots (North Carolina data):
- Finding: mugshots can reliably predict judges’ jail-or-bail rulings; people who appear well groomed or who have wider or rounder faces are more likely to be released.
Medical application example:
- An AI experiment compared electrocardiograms of people who died of sudden cardiac arrest with similar ECGs of survivors; algorithm detected minuscule differences that doctors missed, potentially identifying candidates for pacemakers.
Core insight: algorithms can spot "implausible connections" and operate at a scale and tediousness beyond unaided human cognition.

Returned to MIT in 2024 as professor in economics and electrical engineering and computer science.
Leading initiative: "The Bike Shop @ MIT" — building "bicycles for the mind" (algorithms that amplify human cognition).
- Origin of metaphor: graphic in the March 1973 issue of Scientific American showing "Man on a bicycle" as most efficient animal motion.
Education experiment in India (with Ashesh Rambachan and collaborators):
- Compiling thousands of examples of students’ mathematics homework.
- Goal: use AI to map a "cartography of confusion" to identify where students go wrong and help teachers better diagnose and remediate misunderstandings.
Vision: algorithms as "the new factory floor of science" that can stitch patchwork models together and help transform philosophical questions into definitive science in the next 20 years.

Age reported: 52 years old.
Award and service: won a $500,000 MacArthur Foundation grant in 2002; served on MacArthur Foundation board for 12 years until last June.
Personal interests and quirks: extensive research into ice cream and espresso; led a two-hour tour of Chicago’s ice cream sandwich shops; research into nutrition and exercise.
Described by peers:
- David Laibson: "bridging economics from what we did in the last century to what we will be doing in the next century."
- Lawrence Katz: "His work is as much cited as anyone’s" at a comparable career point.
- Jens Ludwig: described the AI work on bail decisions as "a behavioral economics revolution."

Age: 52 years old.
MacArthur grant: $500,000 in 2002.
Arrest database size (NYC study): more than 700,000 people arrested between 2008 and 2013.
Estimated impacts of risk-assessing algorithm:
- Reduce crime by 25 percent with no change in jail population.
- Reduce jail populations by 42 percent with no increase in crime.
Citation metric: body of work cited almost 100,000 times.
Citation benchmark: 1,000 citations as a field benchmark for influence.
Year of doctorate completion: 1998.
Return to MIT: 2024.
Book publication year: 2013.
Graphic referenced: March 1973 issue of Scientific American.

Sendhil Mullainathan: The AI Economist — Bob Simison, F&D Magazine, December 2025.

The AI Economist
- The AI Economist (Markdown version)
- The AI Economist (PDF)