Engineering Economist Susan Athey Adds DOJ Function to Her Multidimensional Career

After profitable awards for complex advancements in financial concept and industrial organization, serving as chief economist at Microsoft, conducting first exploration that brings together machine finding out with econometric modeling, revolutionary the new discipline of tech economics, helping to identified and foster Stanford HAI, and launching the Golub Money Social Affect Lab at Stanford, Susan Athey will now try out on a new hat as chief economist at the antitrust division of the U.S. Division of Justice (DOJ).

The huge-ranging nature of her operate and occupation has garnered her respect across various academic fields. “Susan is a power of character. She moves from machine finding out to small business tactic to technological innovation coverage to social impact, generating deep suggestions at every single flip,” states Jonathan Levin, the Philip H. Knight Professor and Dean at Stanford Graduate Faculty of Enterprise, exactly where Athey is the endowed Economics of Engineering Professor.

Whilst Athey will keep on her GSB appointment in a component-time potential as she measures into her new authorities role, the shift in her concentration presents an chance to mirror on the significant impression she has had during her vocation and even though at Stanford. 

A Purpose Design for HAI

Couple scientists superior exemplify the multidisciplinary mindset fostered by Stanford HAI than Athey. Her propensity for immersing herself in numerous fields of examine goes back to her undergraduate days at Duke University, the place she graduated in 1991 with a triple-main in economics, arithmetic, and computer science.

Right after earning her Ph.D at Stanford GSB, she went on to come to be a professor of economics and enterprise at MIT, Harvard University, and then at Stanford setting up in 2013, but even within just the area of economics, Athey’s interests have been numerous: In 2007, she received the prestigious John Bates Clark Medal for her contributions to a number of subfields, including industrial business, microeconomic idea, and econometrics.

But it was all through a leave from academia to serve as Microsoft’s chief economist from 2008 to 2013 that Athey created a astonishing link amongst her passion for economics and the tools of AI and machine learning. 

Athey by now knew that digitization and tech platforms were going to perform a substantial function in the economic climate, and that look for engines were poised to have an outsized great importance. She also realized that the analysis neighborhood was just beginning to tackle issues about how to style electronic marketplaces and what healthier competition seemed like in those people markets, and she was enthusiastic to enable build that investigate.

But when she began functioning at Microsoft, Athey also found some thing she was not expecting: the probable for equipment understanding to tackle financial issues. The creators of the Bing lookup engine have been conducting experiments in a method that economists only dreamed about. They were being at the same time operating countless numbers of randomized A/B assessments – asking big quantities of “what if” inquiries to improved have an understanding of these kinds of issues as which look for success need to rise to the leading and how to run auctions for location advertising rates on a lookup web page. By comparison, she says, economists would typically operate a single experiment in a year. 

“Microsoft was applying an synthetic intelligence technique composed of hundreds of algorithms that had been all functioning alongside one another to make a lookup success site,” she claims. “That was a thing new.”

In the field of economics until then, knowledge mining and device discovering had been pejorative phrases for a considerably less highly developed form of stats. “They ended up witnessed as a mechanical training to separate cats from canine,” she says. But at Microsoft, Athey noticed an option to incorporate the computational advancements from predictive device studying with statistical concept so that scientists could better realize causal results not only in organization applications like the lookup engine, but also in social science and economics. It was an epiphany that introduced her in a new analysis route and assisted determine her as one of the early tech economists.

Machine Learning and Causal Effects

Coming out of her encounter at Microsoft, Athey recognized that the insights from predictive algorithms could be harnessed in new techniques by combining them with recent developments in econometrics and studies. For instance, machine studying algorithms could be tailor-made to remedy bring about-and-outcome questions in economics, this kind of as what will happen if we change the least wage? Increase immigration policy? Elevate price ranges? Allow for two corporations to merge? “Predictive equipment learning simply cannot fix these questions by itself, but it can enable,” she says.

For illustration, Athey has utilized equipment understanding to look at the influence on buyers of customized pricing, a sort of cost discrimination that entails charging various costs to consumers according to their willingness to pay. Common economics approaches would give aggregate answers to that trouble, she suggests. They would analyze perhaps just one item classification at a time, looking at demand for, say, distinctive makes of yogurt or towels. By applying device discovering procedures to consumers’ historic buy info, Athey’s study group can estimate customized customer choices across many solutions at the exact time.

Setting up these predictive versions of customer option in convert makes it possible for researchers to talk to even greater questions about these types of things as what happens to prices if you use a tariff, or if generics arrive on the market. “As an enter to answering these questions, we want to have an understanding of how buyers make possibilities,” Athey claims. And equipment studying delivers that input in a way that lets scientists to do this operate more effectively, at a bigger scale, and in a extra individualized way. “If you’re assuming everybody’s the very same, that provides distinctive answers than if you presume that men and women have unique tastes,” she says.

A Tech Economics Pioneer

Athey’s main economist job at Microsoft finished in 2013, but her tenure there described her as one of the first people to be deemed a “tech economist.” It’s a field she has because served build as an independent willpower by convening early conferences in the field and mentoring a lot of pupils together that job path.

“Now tech economists hold an yearly convention that attracts about 800 individuals,” she says. “And we have a specialised occupation market mainly because currently being a tech economist is a distinct job that persons can pursue.”

Athey has also created about what it implies to be a tech economist. “It’s partly a vocation, but it is also a mixture of unique fields of analyze,” she claims. Tech economists study the impression of digitization on the economic system, which involves contemplating about current market layout, privacy, information protection, fairness, competitors policy, and far more, she says. “They also help produce and assess business enterprise versions and aggressive strategy, and they hook up the designs to facts to guideline decisions.” 

Advancing AI for Very good

At Microsoft, in addition to taking an unforeseen deep dive into device discovering and AI, Athey witnessed firsthand the difficulties developed by these technologies – ethical and lawful troubles, To start with Amendment problems, fairness and bias, privateness and copyright, and the prevalence of unforeseen outcomes as folks manipulated or gamed the method in response to market place shifts or new principles.

Mainly because of these observations, Athey produced a desire to impression the ways that machine studying and AI would perform out in the environment. When she returned to academia full time, her initial measures in that route bundled encouraging to system the launch of HAI and then turning into one of HAI’s founding affiliate directors. “Stanford HAI was truly produced to handle these problems,” she suggests. “We want to make AI valuable for humans, and we want to stay away from all of these unintended penalties.”

Athey also wanted to translate the effective uses of AI from the for-revenue sector into the social impact sector. This urge led her to start the Golub Capital Social Impact Lab at Stanford. “We’re bringing the tech toolkit to social affect programs,” she states. So, for instance, the Social Effects Lab has done situation reports of digital training know-how to enhance students’ studying produced and executed ways to targeting instructional messaging to maximize engagement of farmers created and evaluated digital tablet purposes that information nurses by way of counseling people and created procedures to prioritize candidates for scientific trials of COVID-19 remedies.

Connecting the Dots to the DOJ 

Making use of device finding out to attention-grabbing social challenges at the Golub Capital Social Impact Lab is a bottoms-up technique to generating improve, Athey suggests. By distinction, in her new work as main economist of the antitrust division at the DOJ, she will be seeking her hand at addressing the troubles of the digital financial state from the top rated down. “Government guidelines and procedures have an impact on every thing from how competition functions to what mergers go as a result of, to what investments persons make,” Athey suggests.

In going to the DOJ, Athey hopes to carry on several of HAI’s attempts to support governments adapt to an era of promptly shifting technology, particularly close to the use of knowledge in marketplace and in govt. “Because know-how these as artificial intelligence moves so immediately, it is difficult for the authorities to continue to keep up,” she claims. “We have to determine out how all branches of govt are likely to be well prepared to information us through a diverse age.”

It is a great time for Athey to attempt her hand at authorities work, suggests Levin. “At a minute when know-how is ascendant and marketing levels of competition is essential, I can not consider of any person I’d fairly have at the DOJ than Susan.”

Athey will continue to be a professor at Stanford Graduate College of Business and a senior fellow at the Stanford Institute for Financial Policy Exploration. She wunwell phase down from her Stanford HAI affiliate director part while she will carry on to be an affiliated college member.

Stanford HAI’s mission is to advance AI exploration, education, coverage and observe to improve the human situation. Find out a lot more.