Abstract

     Explosive growth in big data technologies and artificial intelligence (AI) applications have led to increasing pervasiveness of information facets and a rapidly growing array of information representations. Information facets, such as equivocality and veracity, can dominate and significantly influence human perceptions of information and consequently affect human performance. Extant research in cognitive fit, which preceded the big data and AI era, focused on the effects of aligning information representation and task on performance, without sufficient consideration to information facets and attendant cognitive challenges. Therefore, there is a compelling need to understand the interplay of these dominant information facets with information representations and tasks, and their influence on human performance. We suggest that artificially intelligent technologies that can adapt information representations to overcome cognitive limitations are necessary for these complex information environments. To this end, we propose and test a novel “Adaptive Cognitive Fit” (ACF) framework that explains the influence of information facets and AI-augmented information representations on human performance. We draw on information processing theory and cognitive dissonance theory to advance the ACF framework and a set of propositions. We empirically validate the ACF propositions with an economic experiment that demonstrates the influence of information facets, and a machine learning simulation that establishes the viability of using AI to improve human performance.

Introduction

     We have begun a conversation with machines that will last for the rest of our lives – that will also be remembered by those machines long after our own fragile memories have failed us.’ William Ammerman 2019

     Big data and artificial intelligence are changing the essence and form of our interactions with information and with information processing machines. No longer are human emotions confined to interpersonal relationships – instead, we find ourselves displaying genuine feelings and sentiment when confronted by intelligent agents and systems (Law, Chita-Tegmark, & Scheutz, 2021). Further, we have witnessed significant increases in information entropy due to social media data of questionable veracity and deliberate attempts to provide misinformation (Moravec, Minas, & Dennis, 2018). As a result, decision makers are confronted by information embedded with greater equivocality, veracity, and density. These developments portend fundamental changes in the nature of information that necessitate attention to the ways in which information is received, conceived, interpreted and acted upon. To this end, we conceptualize and develop the Adaptive Cognitive Fit (ACF) framework, to improve our understanding of how humans can and should leverage artificially intelligent technologies to make decisions in the emerging complex information ecosystems.

Conclusion

     The AI schema posited by ACF can be generalized to a wide range of future information systems and decision support design solutions, where AI can learn from human performance, and environmental variables, to help us in our pursuit of optimality. The projected trajectory of developments in the big data and AI ecosystems, and machine learning in particular “have intensified the speed, and our abilities, to create and deploy new knowledge for constructing theories” (Tremblay et al., 2021, Abbasi et al., 2016). ACF is a unique forward-looking theory, which aligns well with calls from researchers towards positivism (Kar & Dwivedi, 2020) such as the “theory in flux” paradigm (Tremblay et al., 2021): ACF is deeply rooted in prominent theoretical frameworks, and posits a clear application-oriented design science framework that combines AI with HI in its pursuit of optimal performance.

     Information systems discipline has an established culture of creatively theorizing and modeling the interactions between human behaviors, technology and information from an applied and design science perspective (Tremblay et al., 2021, and in general: MISQ Special Issue: Next Generation IS Theories, March 2021). Keeping in line with this valuable culture, we hope that ACF will provide vital insights to ensure the relevance and applicability of cognitive fit research to emerging big data and AI ecosystems.