From Artificial Intelligence (AI) to Intelligence Augmentation (IA): Design Principles, Potential Risks, and Emerging Issues

Authors

Lina Zhou, University of North Carolina at CharlotteFollow
Cynthia Rudin, Duke UniversityFollow
Matthew Gombolay, Georgia Institute of TechnologyFollow
Jim Spohrer, ISSIP.orgFollow
Michelle Zhou, JujiFollow
Souren Paul, Northern Kentucky UniversityFollow

Abstract

We typically think of artificial intelligence (AI) as focusing on empowering machines with human capabilities so that they can function on their own, but, in truth, much of AI focuses on intelligence augmentation (IA), which is to augment human capabilities. We propose a framework for designing intelligent augmentation (IA) systems and it addresses six central questions about IA: why, what, who/whom, how, when, and where. To address the how aspect, we introduce four guiding principles: simplification, interpretability, human-centeredness, and ethics. The what aspect includes an IA architecture that goes beyond the direct interactions between humans and machines by introducing their indirect relationships through data and domain. The architecture also points to the directions for operationalizing the IA design simplification principle. We further identify some potential risks and emerging issues in IA design and development to suggest new questions for future IA research and to foster its positive impact on humanity.

DOI

10.17705/1thci.00185

Recommended Citation

Zhou, L., Rudin, C., Gombolay, M., Spohrer, J., Zhou, M., & Paul, S. (2023). From Artificial Intelligence (AI) to Intelligence Augmentation (IA): Design Principles, Potential Risks, and Emerging Issues. AIS Transactions on Human-Computer Interaction, 15(1), 111-135. https://doi.org/10.17705/1thci.00185
DOI: 10.17705/1thci.00185