Answers are grounded in this paper and cite their sources.
Methodological case study demonstrates improved data triangulation across 212 country cases, highlighting the necessity of structured human-in-the-loop protocols for generative AI use.
The rapid evolution of generative artificial intelligence (AI) is substantially changing archival research methodologies in social sciences. Yet, more structured guidance of methodological protocols informing AI use is needed to mitigate emergent threats to validity, particularly in human rights policy. This article offers a methodological case study, combining retrospective documentation of workflow evolution with descriptive analysis of 417 AI-assisted research sessions within the database supplementing standardized data searches of acid-attack violence for 212 WHO-reporting country cases. Data collection analyzed publicly available digital sources and applied systematic review standards. AI use was slowly streamlined in the last phases of the project. The process of blending AI assistants into the archival research required extensive workflow modifications, including pilot-testing and developing specific research protocols for improved AI use. Protocols include standardizing AI version/model selection and prompts, documenting all sessions, and multi-stage auditing. Benefits included improved data triangulation and more extensive archival searches, particularly in synthesizing non-English sources or when traditional searches faced data scarcity. AI assistants can be a useful supplementary assistant in archival research. Yet, researchers must increasingly explore effective strategies to identify and address methodological concerns, including human-in-the-loop protocols.
No takes yet. Share an insight, caveat, or question.
Jessi Hanson-DeFusco (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: