Paper Type
Complete
Paper Number
PACIS2026-1019
Description
Artificial Intelligence (AI) systems are increasingly driving high-stakes decisions, yet bias can accumulate undetected across every stage of development. Organisations lack a systematic framework for identifying how biases emerge, interact, and accumulate across the AI production pipeline. This study addresses this gap by developing an integrated AI fairness lifecycle framework. First, the paper synthesises prior lifecycle models to develop a five‑stage AI lifecycle integrating business, technical, and deployment perspectives. Second, it introduces a structured typology of bias mitigation strategies comprising data, process, algorithmic, and human‑centred design strategies. Third, it maps bias sources and mitigation strategies across lifecycle stages. Finally, a research agenda is proposed to guide future information systems (IS) research on AI fairness, offering both theoretical and practical guidance for researchers, practitioners and policymakers seeking equitable AI development.
Recommended Citation
Ahuja, Manju; Wang, Jiajun; LEONG, Carmen; and Kuan, Kevin K.Y., "A Lifecycle Framework of AI Fairness and Research Agenda in Information Systems" (2026). PACIS 2026 Proceedings. 1.
https://aisel.aisnet.org/pacis2026/general_topic/general_topic/1
A Lifecycle Framework of AI Fairness and Research Agenda in Information Systems
Artificial Intelligence (AI) systems are increasingly driving high-stakes decisions, yet bias can accumulate undetected across every stage of development. Organisations lack a systematic framework for identifying how biases emerge, interact, and accumulate across the AI production pipeline. This study addresses this gap by developing an integrated AI fairness lifecycle framework. First, the paper synthesises prior lifecycle models to develop a five‑stage AI lifecycle integrating business, technical, and deployment perspectives. Second, it introduces a structured typology of bias mitigation strategies comprising data, process, algorithmic, and human‑centred design strategies. Third, it maps bias sources and mitigation strategies across lifecycle stages. Finally, a research agenda is proposed to guide future information systems (IS) research on AI fairness, offering both theoretical and practical guidance for researchers, practitioners and policymakers seeking equitable AI development.
Comments
17-General