AI-based digital disinformation: A theory-informed and integrated trilateral framework of digital disinformation for information systems research

Digital disinformation poses an increasing challenge to democratic processes, personal decision-making, and business models of digital media platforms. Research has acknowledged this development, examining different types of disinformation, channels of distribution, and dissemination mechanisms. However, prior research gives theoretical insight mainly into the psychological mechanisms on the human-based behavioral level, but lacks theoretical understanding of the increasingly AI-driven nature and process of digital disinformation, as well as of the dynamic socio-technical interplay between human- and AI-driven mechanisms. Drawing on communication theory and disinformation research, this study develops a theory-informed and integrated trilateral process framework, mapping out the essential elements in the emergence, dissemination, and impact of digital disinformation. Our findings suggest three different paths of digital disinformation: human-based, AI machine-based, and hybrid-based, all of which are underlaid by five process layers. The study contributes to our theoretical understanding of digital disinformation and provides novel implications for theory development. Beyond stimulating future IS research, its framework provides a structural integrative process theory for understanding the hybrid, increasingly AI-driven nature of digital disinformation and for bridging the gap between behavioral and technical IS research, allowing thus a more holistic, socio-technical understanding of the disinformation process. The framework serves as a systematic starting point from which to design targeted socio-technical interventions along the disinformation process. • The study provides a theory-informed framework of AI-based disinformation. • It explains the emergence, dissemination and impact of digital disinformation. • There are two distinct paths of digital disinformation: human-based and AI-based. • The study develops novel theoretical propositions for future IS research. • Effective governance of the hybrid and increasingly AI-driven nature of digital disinformation must be socio-technical and multidisciplinary.