HCII Ph.D. Thesis Proposal: Pranav Khadpe
When
-
Description
Perspective-Getting Technologies for Digitally-Mediated Relationships
Pranav Khadpe
HCII Ph.D. Thesis Proposal
Time and Location:
Monday, August 26, 2024 @ 1:00 pm EST
NSH 3305
Meeting ID: 993 8301 7507
Passcode: 933185
Passcode: 933185
Committee:
Chinmay Kulkarni (Co-Chair), Emory University
Geoff Kaufman (Co-Chair), Carnegie Mellon University
Jason Hong, Carnegie Mellon University
Karrie Karahalios, University of Illinois at Urbana-Champaign
Abstract:
What is "appropriate" or "normal"? What is "funny", "offensive", or "disturbing"? What is a "safe assumption"? Too often we have miscalibrated expectations about how others will respond to our actions, and especially when we interact with each other online, where we have fewer cues about our audience. So, when we navigate our relationships online and offline, we can often end up sharing too little or too much. That is, miscalibrated expectations can cause us to underengage in desirable behaviors (fewer meaningful connections) as well as overengage in undesirable behaviors (interactions we regret).
This dissertation explores the concept of perspective-getting affordances in social technologies. Drawing on literature in Social Psychology, I argue that miscalibrated expectations in how others might respond to our actions often stem from egocentric social cognition: our ability to take perspectives is fallible. There is a limit to the conclusions we can draw by simply utilizing our existing knowledge about other people. This egocentric bias can lead to under- or over-engagement in other-oriented behaviors, affecting the quality of social interactions. For instance, people hesitate to initiate interactions because they overestimate the risks of reaching out, thereby foregoing the potential benefits of an enjoyable conversation and more meaningful connection. Similarly, people refrain from addressing divisive topics and conflict because they overestimate how negatively others will respond.
To mitigate these issues, I propose designing social technologies that allow users to get others' perspectives before taking action---to access new information about those they are interacting with, prior to or in the course of taking an action. The research focuses on three main affordances: surfaces for construal alignment, irreducible perspective aggregation, and contingent interactions. These affordances are intended to reduce the perceptual gaps between individuals and enable more accurate social judgments.
The work presents a series of systems, including Hug Reports, Empathosphere, and Nooks, each of which embodies these affordances in the context of professional peer groups. Hug Reports supports exchanges of appreciation among developers, Empathosphere supports open communication in ad-hoc virtual groups, and Nooks supports initial interactions among workplace peers. Each of these systems were evaluated through either a field deployment or controlled study, demonstrating their potential to improve social interactions by reducing egocentric bias.
My proposed work aims to examine the implications of AI-driven perspective-getting tools, exploring how AI might assist in generating and interpreting others' perspectives in social contexts. The findings contribute to the design of social technologies that not only facilitate perspective-getting but also support healthier and more effective social interactions in digitally-mediated relationships.
This dissertation explores the concept of perspective-getting affordances in social technologies. Drawing on literature in Social Psychology, I argue that miscalibrated expectations in how others might respond to our actions often stem from egocentric social cognition: our ability to take perspectives is fallible. There is a limit to the conclusions we can draw by simply utilizing our existing knowledge about other people. This egocentric bias can lead to under- or over-engagement in other-oriented behaviors, affecting the quality of social interactions. For instance, people hesitate to initiate interactions because they overestimate the risks of reaching out, thereby foregoing the potential benefits of an enjoyable conversation and more meaningful connection. Similarly, people refrain from addressing divisive topics and conflict because they overestimate how negatively others will respond.
To mitigate these issues, I propose designing social technologies that allow users to get others' perspectives before taking action---to access new information about those they are interacting with, prior to or in the course of taking an action. The research focuses on three main affordances: surfaces for construal alignment, irreducible perspective aggregation, and contingent interactions. These affordances are intended to reduce the perceptual gaps between individuals and enable more accurate social judgments.
The work presents a series of systems, including Hug Reports, Empathosphere, and Nooks, each of which embodies these affordances in the context of professional peer groups. Hug Reports supports exchanges of appreciation among developers, Empathosphere supports open communication in ad-hoc virtual groups, and Nooks supports initial interactions among workplace peers. Each of these systems were evaluated through either a field deployment or controlled study, demonstrating their potential to improve social interactions by reducing egocentric bias.
My proposed work aims to examine the implications of AI-driven perspective-getting tools, exploring how AI might assist in generating and interpreting others' perspectives in social contexts. The findings contribute to the design of social technologies that not only facilitate perspective-getting but also support healthier and more effective social interactions in digitally-mediated relationships.
Proposal document: