Blame Ascriptions Toward Autonomous Agents

As autonomous agents become embedded in organizational and everyday contexts, humans are increasingly delegating tasks to them. Yet when undesirable outcomes occur, people may blame the autonomous agent. This thesis investigates how blame is socially expressed, referred to as blame ascriptions, toward autonomous agents following IS delegation and undesirable outcomes . I present the results of a three-phased literature review in Chapter 2. Drawing from multiple theoretical perspectives, I develop a conceptual model and theoretical propositions in Chapter 3 that organize the process of ascribing blame toward autonomous agents into Triggering Blame, Enacting Blame, and Diagnosing Blame. Using a theoretical typology, I posit three unique blame ascriptions toward autonomous agents: Diminished Control, Ignorance, and Condemnation. Chapter 4 presents the results of a qualitative study involving 34 interviews across three domains, where I validate the proposed narratives and uncover a fourth narrative called Erosion of Skills. These narratives reveal that humans ascribe blame toward autonomous agents for superseding human control, confounding human judgment, committing moral harm, or diminishing human skills. Chapter 5 presents the results of a quantitative study, which uses an experimental vignette design to examine how the perceived agency attributes of autonomous agents influence blame ascription and IS delegation. The findings show that the perceived agency attributes of autonomy and inscrutability lead to increased blame ascription, blame ascription is negatively associated with IS delegation, and reflection may reduce the ascription of blame. This research contributes to information systems scholarship by theorizing blame ascription as a socially constructed and narratively expressed response to undesirable outcomes perceived to be caused by autonomous agents. Contrary to prevailing research that emphasizes the negative implications of blame, this thesis demonstrates that certain forms of blame ascriptions can serve a productive role in diagnosing outcomes and guiding future design and use practices. Moreover, the results of the two studies suggest the importance of context in the ascription of blame, and that reflection may play a pivotal role in the relationship between the enactment of a blame ascription and future IS delegation to autonomous agents.

Department dc:contributor.department

Business

Year dc:date.issued

2025

Author dc:creator

• Killoran, Jayson Andrew

Advisor dc:contributor.supervisor

• Brohman, Kathryn

• Blame ascription
• IS delegation
• Perceived agency
• Undesirable outcome
• Autonomous agent

Statement dc:rights

• Attribution 4.0 International

Licence dc:rights.uri

• http://creativecommons.org/licenses/by/4.0/

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/1974/34667

OAI identifier oai:identifier

oai:queensu.scholaris.ca:1974/34667