Abstract
As Generative Artificial Intelligence (GAI) increasingly automates and/or augments knowledge- and creativity-intensive tasks, what does this mean for knowledge work? We investigate its implications through an 18-month case study of a multinational firm, involving data scientists, machine learning experts, software engineers, and other information technology professionals who employ GAI in their daily work. Information systems research often investigates the effects of Artificial Intelligence through the lens of automation or augmentation, yet it overlooks task-level variation and workflow interdependencies and oversimplifies GAI’s impacts on knowledge work. Adopting a task-oriented perspective, we extend the notion of automation-augmentation by specifying the dynamics between the two and demonstrating that their effects vary according to task characteristics and task interdependencies within role-specific bundles. We theorize that GAI reconfigures knowledge work through four patterns: augmentation with automation, augmentation without automation, automation as task obsolescence, and augmentation as task invention. We then show how these effects propagate across interconnected tasks and accumulate into changes in governance, roles and expertise, and learning approaches. We offer a set of propositions that advance understanding of human-technology relationships within knowledge work and translate our findings into practical implications for the future of work.
DOI
10.17705/1jais.01023
Recommended Citation
Jiang, Mingyuan; Karanasios, Stan; and Breidbach, Christoph, "Where Generative AI Automates, Where It Augments: Reconfiguration Patterns in Knowledge Work" (2026). JAIS Preprints (Forthcoming). 258.
DOI: 10.17705/1jais.01023
Available at:
https://aisel.aisnet.org/jais_preprints/258