Positioning Paradata: A Conceptual Frame for AI Processual Documentation in Archives and Recordkeeping Contexts
Scott Cameron, Patricia C. Franks, Babak Hamidzadeh
Abstract
The emergence of sophisticated Artificial Intelligence (AI) and machine learning tools poses a challenge to archives and records professionals, who are accustomed to understanding and documenting the activities of human agents rather than the often-opaque processes of sophisticated AI functioning. Preliminary work has proposed the term paradata to describe the unique documentation needs that emerge for archivists using AI tools to process records in their collections. For the purposes of archivists working with AI, paradata is conceptualized here as information recorded and preserved about records’ processing with AI tools; it is a category of data that is defined both by its relationship with other datasets and by the documentary purpose it serves. This article surveys relevant literature across three contexts to scope the relevant scholarship that archivists may draw upon to develop appropriate AI documentation practices. From the statistical social sciences and the visual heritage fields, the article discusses existing definitions of paradata and its ambiguous, often contextually dependent relationship with existing metadata categories. Approaching the problem from a sociotechnical perspective, literature on Explainable Artificial Intelligence (XAI) insists pointedly that explainability be attuned to specific users’ stated needs—needs that archivists may better articulate using the framework of paradata. Most importantly, the article situates AI as a challenge to accountability, transparency, and impartiality in archives by introducing an unfamiliar non-human agency, one that pushes the limits of existing archival practice and demands the development of new concepts and vocabularies to shape future technological and methodological developments in archives.