Date: 15-09-2021
About the September 15th talk: The behavioural revolution in social sciences, especially economics, was essentially an empirical revolution inspired by psychology and aided by experimental methods. It challenged our understanding of human decision making by modifying the rational choice approach of neo-classical economics and discovering that human beings are far less rational than formerly assumed. Hitherto, the rational choice approach mistakenly assumed that the mere ability to be rational was evidence of a consistent propensity to be rational – much like inferring that humans run all the time because they have the occasional capacity to run. The field of AI arose in its early stages as an attempt to mimic human intelligence using computers. It made vast strides in recent years based on cheap digital memory and processing capacity leading to vast amounts of data about people’s choices being collected. The methodology of popular, commercially applicable AI (used interchangeably with ML) relies on a couple of statistical methodologies that seek to further prediction and clustering.
Both fields are tantalizingly close and occasionally intersect. Behavioural sciences promise to tell us more about actual human behaviour, informing what AI has to emulate. In turn, AI can help as a methodological tool in understanding human behaviour much better than traditional research methods (qualitative, observational and experimental). The talk will cover the following theses representing two directions of influence:
A) Behavioural Science -> AI
Emulating irrational behaviour – makes for more realistic AI;
B) AI -> Behavioural Science
AI helps build more heterogeneous models of human behaviour, involving more covariates than experiments, while at the same time avoiding the overfitting to which the more context-driven social sciences, such as sociology and anthropology, regress. AI can help Behavioural Sciences overcome the temptation of building long, contingent lists of context-specific behaviours, thereby rescuing theory from particularit
Speaker Bio: Pavan Mamidi is the Director of the Centre for Social and Behavior Change (CSBC) at Ashoka University. He helped set up India’s first Nudge Unit at Niti Aayog in New Delhi, and is currently helping set up a Nudge Unit in Uttar Pradesh as well. Previously, he ran a behavioral experiments lab for the University of Oxford in India, and also worked at IIM Ahmedabad, IIM Bangalore, University of Michigan Law School, MIT Sloan, and Harvard Law School. His areas of research interest cover social norms, extralegal behavior, and behavioral ethics. He has a Doctorate from Oxford, a Masters from Harvard Law School, and a bachelors in Mathematics from Osmania University.
Dr. Mamidi’s talk will be followed by a discussion and Q&A session anchored by Dr. Mor.
CITAPP’s Monthly Seminar Series is an attempt to create a forum where researchers across IIITB domains can meet and discuss cutting edge research on the chosen theme of the semester. The Series hopes to explore a technology or topic for its ramifications in different realms of social activity. In particular, we are interested in understanding the specific kinds of complexity that these domains present for technological innovation and design.