Audited AI-assisted archival research on acid-attack violence

A methodological case study describes AI assistants supplementing archival searches on acid-attack violence, with standardized model and prompt choices, session logging and multi-stage auditing. The author reports descriptive analysis of 417 sessions across 212 WHO-reporting country cases.

Workflow level n/a Source Peer-reviewed source Outcome Author-reported

Evidence & verification

Source level

Peer-reviewed source
Grounded in a peer-reviewed publication. This describes the source; it does not mean the reported results have been independently reproduced.

Outcome verification

Author-reported
Outcomes are reported by the team that built the system. The entry is source-checked against the team's own publications; not independently reproduced.

Sources checked

2026-10-17 (UTC)

abstract-only primary-source AI editorial review; source verification is separate from outcome verification.

Source-checked does not mean independently reproduced. See the Method page for how these labels are assigned.

Sources (1)

Research applicability

Accept a bounded social-science archival workflow supported by the primary abstract. Benefits in triangulation and non-English searching remain author-reported, not a measured editorial comparison. Session and country-case counts describe the study, not verified prevalence or universal geographic coverage.

Getting started

Editorial, untested adoption guidance: read the article, define a bounded archival question, fix and record model/prompt versions, and check every proposed source and extraction before analysis.

Maintenance

Paper-based workflow; ongoing software maintenance not established.

License

Reuse rights for text/data/software must be checked separately; publication access does not establish unrestricted reuse.

Workflow & human oversight

Researcher-led workflow: people set research questions, approve data access and adjudicate outputs; no autonomous end-to-end performance claim. Workflow level intentionally unspecified.

Workflow level is not an evidence score; higher is not better evidence.

Approach

Limitations

Abstract-only publisher review; extraction returned an incomplete introduction, which is not used to infer methods. No exact model versions, prompts, audit error rates, dataset access or executable recipe verified. Retrospective descriptive case study, not independent replication. Sensitive violence-related records require rights, privacy and contextual checks.

Reproducibility

No software executed, participants studied or reported outcome independently reproduced by this reviewer.

Adoption requirements

Access rights to archival sources, model access, session/provenance logging and domain-informed researchers able to audit multilingual evidence. Do not upload identifiable survivor information without an appropriate lawful and ethical basis.

Comments

No comments yet.

Add a comment

Comments are held for moderation before appearing. Critical or corrective comments are welcome — especially evidence that an entry overstates what has been verified.