Leveraging artificial intelligence for ethical archiving and democratising access to sensitive historical narratives.

Robust historical documentation is foundational to post-conflict transitional justice efforts, enabling accountability, reconciliation, and societal healing. Zimbabwe's experience following the Gukurahundi period (1983-1987), a conflict involving significant civilian casualties, particularly among Ndebele communities in Matabeleland and the Midlands, highlights critical challenges in archiving sensitive histories. Constraints in accessing fragmented, incomplete, or vulnerable records, alongside limited participatory frameworks for historical testimony, impede inclusive truth-telling and community engagement in national healing processes. Artificial intelligence (AI) offers transformative potential to address these archival limitations by enabling large-scale data aggregation, automated anonymisation of sensitive testimonies, and secure decentralised storage. Such methodologies democratise historical preservation, ensuring ethical documentation even in contexts where traditional archival systems face structural challenges. As such, this study sought to address Zimbabwe's archival asymmetry by co-designing, with stakeholders, a prototype digital AI-driven archive that prioritises data sovereignty and intersectional inclusivity. A transdisciplinary and participatory action research (PAR) approach structured across three key phases, which include stakeholder mapping and needs assessment, co-design workshops, and prototype development and testing, was employed. The study was the first in Zimbabwe to build a co-designed, AI-driven digital archive specifically tailored to a politically charged historical context, offering a new paradigm for ethically managing post-conflict records. [ABSTRACT FROM AUTHOR]