Emory University’s Department of Economics hosted an event about artificial intelligence’s impact on growth, jobs and inequality on Oct. 27. The event featured Thomas Sargent, the 2011 recipient of the Nobel Memorial Prize in Economic Sciences and William Berkley professor of economics at New York University.
Sargent defined artificial intelligence as two separate phrases: artificial, meaning non-human, and intelligence, meaning human. He explained that human intelligence is characterized by pattern recognition, decision making and an awareness of time and space.
During the event, Sargent identified Steven Pinker, a world renowned psychologist, as a notable contributor to his knowledge. Namely, Sargent said that humans “weren’t evolved to know” biology, statistics and economics, yet these subjects are used to create artificial intelligence.
Extending beyond contemporary figures, Sargent’s intellectual influences also include Galileo Galilei and Charles Darwin. He explained that both used economic tools to generalize conceptions about the world.
“Artificial intelligence and machine learning are descendents of methods Galileo first used,” Sargent said.
Machine learning entails “building a model of the world and using the model to make decisions,” Sargent said. This makes machine learning a key facet to artificial intelligence, since it is part of what supplies artificial intelligence tools with access to data so predictions can be made.
Sargent listed economic topics like statistics, calculus, linear algebra, optimization and linear programming as foundational tools for artificial intelligence.
“Lots of ideas from economics have been imported into artificial intelligence,” Sargent said.
Multi-agent decision theory is an important aspect of economics, which can produce research into a realistic approach to social structure modeling, modeling the market and predicting an appropriate response to a real-life crisis. This theory was computerized into a form of artificial intelligence that can help economists make predictions.
Sargent shared two major related triumphs involving the intersection of machine learning and economics: an infamous chess match and a game of Go.
During a 1997 chess match, the top ranked chess player in the world, Garry Kasparov, played against a supercomputer called Deep Blue, andor the first time, artificial intelligence beat a human. This was a monumental feat and helped people begin to see a future for this field.
In 2017 Google’s DeepMind artificial intelligence beat the No. 1 Go player in the world, Ke Jie. Go is regarded as one of the most complex game strategy board games, so Sargent said this victory marked a new frontier for artificial intelligence.
A Q&A period followed Sargent’s lecture and attendees conveyed their fear about artificial intelligence replacing jobs conducted by humans.
Sargent responded that although there have been many jobs where artificial intelligence has outcompeted humans, there are a plethora of new jobs created to manage the technology.
“One thing you can do in education is not get stuck. Plan for flexibility and some robustness,” Sargent said. “I think there are a lot of jobs that are going to get better because of artificial intelligence and machine learning.”
Another attendee asked about the real life applications of artificial intelligence. Sargent explained that one form of artificial intelligence is used for second-price auctions to quickly allocate ads. He noted that even the IRS uses machine learning to fight tax return lobbies.
Closing out the seminar, an attendee asked Sargent for advice on pursuing a career in economics. He advised that students take “linear algebra, calculus and statistics,” but more importantly, suggested that they “persist.”
