Human-Computer Interaction Ph.D. Thesis Proposal
Speaker
LYNN KIRABO
Ph.D. Student
Human-Computer Interaction Institute
Carnegie Mellon University
When
-
Where
In Person and Virtual - ET
Description
Public transit (PT) is the heartbeat of most cities around the world. It gives communities access to employment and services like health and education. Policy recommendations, interventions, and research on PT often focuses on drivers as the primary stakeholder. This same focus is evident in the recent proliferation of machine learning (ML) interventions in PT technologies. They neglect the influence and impact of these ML interventions on other stakeholders in the PT ecosystem. This focus runs the risk of automating inequities within future mobility systems.
In this dissertation, I argue that to design for equity in public transit, we should have an understanding of the broader PT ecosystems in which we are deploying transit AI technologies. My completed research studies traverse two geographic contexts, East Africa and North America. My work shows an underlying influence of trust on relationships within the ecosystem, and unique stakeholder appropriation of transit technologies. Conversely, we also found a suspicion of advanced smart transit interfaces. Thus, we propose that to design for equity in smart transit systems, designers and researchers should consider two dimensions of trust: trust in the interfaces and trust between stakeholders within the ecosystem. My proposed work focuses on the first dimension, trust in the interface. In Study 5, we will create a taxonomy of XAI principles, essential for the creation of inclusive smart transit interfaces, through co-design activities with disabled riders in North America. In study 6, we plan to demonstrate how these principles can improve users' experiences with smart transit interfaces by testing an interface that embodies the XAI principles. This thesis makes the following major contributions: 1) Establishes a multidimensional connection between Trust and Ecosystems, 2) Demonstrates a need to understand the entire ecosystem when considering new technologies, and 3) Demonstrates how methodologies can be adapted for research in these areas.
Thesis Committee:
Aaron Steinfeld (Chair)
Jessica Hammer
Jeffrey P. Bigham
Destenie Nock
Tawanna Dillahunt (University of Michigan)
Additional Information
In Person and Zoom Participation. See announcement.