The Threatening Effects of Help from Agentic Information Systems and their Implications for User-System Collaboration (Harmful Help from Information Systems)

• Applicant: Professor Dr. Martin Adam
• Subject Area: Operations Management and Computer Science for Business Administration ; Data Management, Data-Intensive Systems, Computer Science Methods in Business Informatics
• Term: since 2023
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 522190413

The results of the first research project are threefold: (1) conceptualizing the nature of help invocation types (HITs) from (agentic) IS, particularly reactive and anticipatory help; (2) identifying the effects of anticipatory (vs. reactive) help from IS on users’ perceptions, particularly self-threats related to competence and identity; and (3) identifying the effects of anticipatory (vs. reactive) help on user behavior and user-IS collaboration outcomes, particularly related to delegation. In light of these scientific results, further research gaps have been identified, particularly in terms of how HITs relate to user reflection and related outcomes. Indeed, it is practically crucial and theoretically intriguing to empirically investigate how users process the IS-provided information to reflect and improve their understanding and decision-making. In this vein, we lack empirical insights into how HITs support or threaten a user’s understanding within the process of reflection. Relatedly, a user cannot only perceive the HIT-provided information as an improvement to the user’s knowledge of a domain, thereby broadening the users’ view of domain and improving related outcomes (e.g., critical reasoning) – but may consider it a perceived threat to the user’s understanding, fortifying the user’s old views (e.g., confirmation bias, “backfire effect”) and leading to minimal changes to the users’ understanding and related outcomes. Against this backdrop, three new and central research questions have been derived, representing the focus of this follow-up research project: (RQ1) What key conceptual dimensions constitute user reflection with (agentic) IS, and how can they be designed and quantitatively measured?; (RQ2) What are the effects of HITs, particularly anticipatory (vs. reactive) help from IS, on users’ distinct states of reflection?; and (RQ3) How do anticipatory (vs. reactive) help affect reflection-related user behavior and user-IS collaboration outcomes (e.g., critical reasoning)? With the follow-up research project, we aim to extend our contribution to the literature on help from (agentic) IS and user-IS collaboration by providing a more nuanced theoretical understanding of HITs from IS and their reflection-related implications and designs, which become increasingly important with the spread of more agentic IS. Given that the project operates at the interface between IS, human-computer interaction, and related fields, we expect our results to influence an audience that extends beyond the IS community.

• DFG Programme: Research Grants