Type: article
Abstract: Synthetic data (SD) is increasingly adopted in domains such as finance, to overcome privacy, scale, and access constraints. Although SD enables experimentation and data sharing, it also raises fundamental questions about the relationship between data and reality. Ironically, this is particularly salient in the same domains, as decisions have far-reaching monetary and socioeconomic consequences. Drawing on 17 interviews with practitioners in banking and finance, this paper examines the drivers, affordances, and challenges of SD. We identify three paradoxes that characterize SD use: scarcity versus validity, fluidity versus anchoring, and fidelity versus drift. From these, we develop a higher-order generativity–grounding paradox, showing how the capacity to generate SD simultaneously weakens the connection to empirical referents. We contribute to the information systems literature by theorizing SD as a paradoxical instantiation of data, shifting the focus from managing the tensions in data use to understanding the tensions embedded in the data itself.
Source: AIS Electronic Library (AISeL)
Authors: Goswami, Suheli, Hekkala, Riitta
Institution: Aalto University
Language: English
Subfield: Management Information Systems
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W7222989208
