Paper Type

Complete

Abstract

This study examines the unintended side effects of AI-based artifacts across multiple domains. Through a systematic literature review, we analyzed recent studies on AI technology and identified 15 side-effect dimensions. These dimensions are organized into a multi-level structure spanning individual, organizational, and societal contexts. The findings show that AI side effects are interconnected consequences of recurring challenges, including overreliance, biased data, opaque systems, and governance limitations. Building on these insights, this study proposes an inductive framework that explains how AI artifacts, usage contexts, and underlying challenges generate side effects that accumulate and propagate across levels. This paper also synthesizes risk-mitigation strategies and outlines a research agenda to guide future studies on the unintended consequences of AI adoption.

Paper Number

1341

Comments

SIG HCI

Download

Share

COinS
 
Aug 15th, 12:00 AM

Side Effects of AI Use: How They Affect Different Stakeholders

This study examines the unintended side effects of AI-based artifacts across multiple domains. Through a systematic literature review, we analyzed recent studies on AI technology and identified 15 side-effect dimensions. These dimensions are organized into a multi-level structure spanning individual, organizational, and societal contexts. The findings show that AI side effects are interconnected consequences of recurring challenges, including overreliance, biased data, opaque systems, and governance limitations. Building on these insights, this study proposes an inductive framework that explains how AI artifacts, usage contexts, and underlying challenges generate side effects that accumulate and propagate across levels. This paper also synthesizes risk-mitigation strategies and outlines a research agenda to guide future studies on the unintended consequences of AI adoption.