Toward Human-Centered Algorithmic Technologies
Data-driven algorithmic technologies increasingly act as mediators between people and the world around them – curating Facebook walls, managing smart city infrastructure, and informing government policy decisions on important issues like immigration law. These technologies can help organizations process massive amounts of data and provide insights beyond what the human eye and mind could perceive alone. Yet despite their often-touted objectivity and efficiency, behind the scenes, these algorithms can make biased, unfair, incomprehensible, and untrustworthy decisions, with unintended consequences for human decision-making and work practices.
In my research, I draw on social science theories and methods to understand the consequences of algorithmic technologies. I use this understanding to design tools and interaction techniques for algorithmic technologies that can accommodate more diverse human motivations, values, and social and decision contexts, in order to improve the ways people live and work.
In this talk, I will present the findings of a qualitative study I conducted with drivers from Uber and Lyft, two ridesharing companies that use algorithms to allocate, incentivize, and evaluate work. My study suggests that while algorithmic management enables efficient, large-scale services to operate with only a few human managers in each city, it fails to account for important dimensions of human motivation and control, and complicates multiple stakeholders’ conflicting interests. These challenges decrease worker cooperation and trust. I approach this problem by examining three interacting components of algorithmic technologies: the mental models and biases that people bring to their interactions with such technologies; the interfaces that are used to communicate algorithmic results; and the logic and working mechanisms of the algorithms themselves. I will discuss my past and ongoing work on each of these components in the areas of health care, smart city technology, and on-demand work.
Min Kyung Lee is a research scientist in human-computer interaction at the Center for Machine Learning and Health at Carnegie Mellon University. Her research examines the social and decision-making implications of intelligent systems and designs tools and interaction techniques to improve decision-making and collaboration. Dr. Lee is a Siebel Scholar and has received several best paper awards, as well as an Allen Newell Award for Research Excellence. Her work has been featured in media outlets such as the New York Times, MIT Technology Review, New Scientist, and CBS. She received a Ph.D. in HCI in 2013 and an M.Des. in interaction design from Carnegie Mellon, and a BS summa cum laude in industrial design from KAIST.