Stanford Economics Professor Susan Athey
The development of artificial intelligence is endogenous. That is, it’s not just some pre-specified innovation path that we’re all going to go down. Instead, it’s something that is responding to incentives. Of course, AI is what it is, it has technological capabilities and limitations, so it only can do so much. But the ways in which we improve it and develop it really depend on incentives.
Professor Susan Athey, The Economics of Technology Professor at the Stanford Graduate School of Business, discussed the trends and future of artificial intelligence (AI) in the social sector during last year’s Graduate School of Business virtual alumni week. She emphasizes that universities can play a role in guiding the development of AI in directions that can be beneficial to humanity.
Below is an overview of Professor Athey’s talk, “AI for Social Good.” Recorded on September 28, 2021, during Stanford’s Graduate School of Business Virtual Alumni Week.
Artificial intelligence (AI) can yield benefits for the social sector. Universities can help guide AI development, setting out best practices to ensure fairness and governmental efficiency.
- AI development is endogenous; that is, it does not follow a pre-specified path, but responds to incentives; policymakers could push AI development toward augmenting human capabilities rather than replacing human workers.
- Ideally, firms could anticipate the unintended future consequences of their technologies, but this is unlikely to happen; furthermore, much of this research should be done by disinterested parties.
- AI creates opportunities to benefit humanity in the social sector, in government, and in the private sector; services such as the retraining of workers could be automated and delivered remotely, maximizing accessibility.
- The private and social incentives to develop AI with widespread benefits might not be aligned; universities should make a push to develop socially beneficial AI applications.
- Firms seek profits and market power.
- Firms might not invest in products that can easily be copied.
- Firms might not invest in products that serve low-income groups, such as educational apps.
- Research shows that trends in AI and automation respond to market conditions; technologists and investors are developing robots in response to labor shortages.
- Firms are investing in automation where the workforce is aging, making labor expensive.
- Over the next twenty years, economists forecast a shortage of workers in the middle stages of their careers.
- Forecasters should consider that demographic changes may be more important than changes due to robots.
- Key advances in AI include iterative experimentation, used in automated recommendation systems; these systems are suitable for use where the cost of errors is low and the environment does not change drastically.
- Use of AI in applications like resume screening can exaggerate bias, or eliminate bias, and firms could benefit from following best practices developed by the social sector; we should guide AI innovation to address concerns about fairness, bias, access, and market power.
- Within large firms, the incremental innovation of new AI applications involves multiple stages, from the collection of data from previous experiments, to advanced analysis of the system’s objectives, to improvements; universities can help governments and other organizations replicate this process.
AI does not evolve along a pre-determined path but responds to incentives. Firms may lack the incentives to research the unintended consequences of their technologies or to develop innovations for certain markets. Universities can ensure that disinterested parties consider the long-term consequences of AI, and develop AI applications for neglected markets. Substantial investment in AI occurs in response to labor shortages; dystopian forecasts about what AI means for the future should consider the importance of demographic trends. The use of AI in applications like resume screening illustrates concerns about AI bias and fairness. Universities might consider how to develop best practices to guide AI innovation, and help governments and the social sector engage in incremental improvements of AI-based applications.
View Professor Athey’s talk, “AI for Social Good.” Recorded on September 28, 2021, during Stanford’s Graduate School of Business Virtual Alumni Week.
Susan Athey, an economic theorist who has made significant contributions to the study of industrial organization, is the Economics of Technology Professor at Stanford Graduate School of Business. She also is Professor of Economics (by courtesy), School of Humanities and Sciences, and Senior Fellow, Stanford Institute for Economic Policy Research. Her current research is in the areas of the economics of digitization, marketplace design, and the intersection of machine learning and econometrics. She has studied a range of application areas, including timber auctions, online advertising, the news media, and the application of technology for social impact.
About Susan Athey
Susan Athey, an economic theorist who has made significant contributions to the study of industrial organization, is the Economics of Technology Professor at Stanford Graduate School of Business. She also is Professor of Economics (by courtesy), School of Humanities and Sciences, and Senior Fellow, Stanford Institute for Economic Policy Research. Her current research focuses on the economics of the Internet, marketplace design, auction theory, the statistical analysis of auction data, and the intersection of computer science and economics.