Abstract
There is significant interest and research in engineering machines, algorithms, and systems that humans will trust. However, serious questions remain about trust: what it is, whether it can be measured and modeled, whether it provides useful information for engineering and design, and whether the engineering of trust will be deployed ethically. This proposed session will facilitate a discussion between leading proponents and skeptics of trust research to help provide insights into these very important questions.
Keywords
Subject Matter
Human-machine, human-robot, and human-automation trust has been the subject of human factors research going back to the 1970s (Bainbridge, 1983; Halpin, Johnson, & Thornberry, 1973; Lee & See, 2004; Montague, Winchester III, & Kleiner, 2010; Parasuraman & Riley, 1997; Sheridan & Ferrell, 1974). Recent history has seen an intensification of trust research, with many contributions being made by investigators both traditionally in human factors disciplines and outside of them. Despite this history, there is still intense debate about trust. Such debates can range from the fundamental to the ethical: what trust is; what factors influence it; whether it can be measured or modeled; whether it satisfies the basic requirements of validity, diagnosticity, and selectivity; whether measured trust ratings are cardinal and thus meaningfully, mathematically modeled; and what the moral and ethical implications are for engineering it (Bolton, 2022; Bolton, Biltekoff, Wei, & Humphrey, 2022; Dekker & Nyce, 2015; Hancock, 2023; Han cock, Billings, & Schaefer, 2011; Hancock, Weaver, & Parasuraman, 2002; Wei, Bolton, & Humphrey, 2020). These issues have made some question whether trust is valuable as an engineering construct at all.
Summary of Objectives
This session will seek to enable an open discussion around these questions. To this end, this session has engaged five experts on trust who varyingly support or are skeptical of it. All are committed to a civil discussion around the assorted views of trust and to debate whether it is a concept that has value in human factors and engineering.
It is the hope of session organizers that that audience will develop a deeper appreciation of the practical, philosophical, and ethical issues around trust. It is also our hope that the discussion will stimulate future thought and debate around issues that will be addressed in the broader human factors community. Finally, given the current popularity of trust research, this is a very timely discussion that will likely be well attended.
Presentation Format
This session will generally follow a panel discussion pattern. However, it is being proposed as an “alternative format” because it intends to be somewhere in between a debate and a traditional panel (it will be a structured discussion around a possibly contentious subject). It also involves potentially complex arguments deserving of proper time for development. This is something that could be difficult with the requisite 5-minute opening statement times of standard discussion panels.
The proposed session will involve 5 participants/panelist and proceed as follows:
(40 minutes maximum) Each participant will have 8 minutes to introduce themselves and present/describe their position on trust and why they do or do not think trust should be an engineered construct.
(20 minutes) Participants will engage in a moderated discussion where they will be able to ask and answer questions between each other based on the preceding presentations.
(25 minutes) Participants will respond to audience questions and comments.
(5 minutes) Participant’s will be able to each make an 1 minute closing statement.
Any time not used during opening presentations will be added to the time allowed for discussion and questions.
Room Requirements
The session will not require any special accommodations beyond what would be needed for a typical lecture section and discussion panel:
A setup for presenting from a computer (e.g., podium, computer, projector(s), audio connection);
Microphones for up to five participants;
One or more microphones for enabling audience questions and comments; and
Seating for the audience.
Participants
All the participants are listed above as co-authors. All have well-established publication records on trust and its surrounding issues. Michael Rayo (Ohio State University) will serve as the session’s moderator/session chair.
