Paper Number
ICIS2025-2543
Paper Type
Short
Abstract
As artificial intelligence systems take on increasingly autonomous roles in human-centered domains, their lack of emotional intelligence has emerged as a critical limitation. Multi-agent systems demonstrate goal-directed autonomy but remain largely indifferent to the affective dynamics that shape trust and collaboration. This paper introduces the EMPACT framework for embedding emotional intelligence into multi-agent architectures. Drawing on theories of emotional intelligence, affective computing, and design oriented IS research, EMPACT reframes emotional reasoning as a constitutive property of agency. The framework comprises five interdependent modules: Self-Awareness, Perception and Empathy, Regulation and Motivation, Social Interaction, and Feedback and Adaptation. We further propose an evaluation roadmap grounded in the DSVF, addressing criterion, causal, context, and normative dimensions. By foregrounding emotional intelligence as central to digital agency, this work contributes to IS debates on responsible, human-centered AI and outlines a pathway for designing systems that are both autonomous and relationally attuned.
Recommended Citation
Tuguldur, Temuulen, "EMPACT Loop: Design Framework for Emotionally Intelligent Multi-Agent Systems" (2025). ICIS 2025 Proceedings. 34.
https://aisel.aisnet.org/icis2025/hti/hti/34
EMPACT Loop: Design Framework for Emotionally Intelligent Multi-Agent Systems
As artificial intelligence systems take on increasingly autonomous roles in human-centered domains, their lack of emotional intelligence has emerged as a critical limitation. Multi-agent systems demonstrate goal-directed autonomy but remain largely indifferent to the affective dynamics that shape trust and collaboration. This paper introduces the EMPACT framework for embedding emotional intelligence into multi-agent architectures. Drawing on theories of emotional intelligence, affective computing, and design oriented IS research, EMPACT reframes emotional reasoning as a constitutive property of agency. The framework comprises five interdependent modules: Self-Awareness, Perception and Empathy, Regulation and Motivation, Social Interaction, and Feedback and Adaptation. We further propose an evaluation roadmap grounded in the DSVF, addressing criterion, causal, context, and normative dimensions. By foregrounding emotional intelligence as central to digital agency, this work contributes to IS debates on responsible, human-centered AI and outlines a pathway for designing systems that are both autonomous and relationally attuned.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
15-Interaction