S.6041 ALGORITHMIC SOLUTIONS TO HUMAN PROBLEMS (SPRING)
S.6041 Algorithmic Solutions to Human Problems: Integrating Artificial Intelligence and Behavioral Science
There are many challenging human problems for which machine learning algorithms could provide a solution. One specific example: Many poor people do not get benefits they are eligible for (e.g. food stamps) because the application process is psychologically overwhelming. If we can build a copilot for coding, can we build a red tape copilot that helps people navigate this process? Building such AI solutions, however, requires more than coding skills and data. It requires a more sophisticated model of machine learning (novel computational frameworks), a deeper understanding of people (insights from behavioral economics), and computational models that integrate the two.