Artificial Intelligence and Research on the Consequences of War: Thematic overview of eight papers and preparation for project implementation

Abstract

This overview synthesises eight papers into a research programme for using artificial intelligence to investigate the environmental consequences of war. It proposes a scientist–AI system in which expert judgement, explicit responsibility, hypothesis generation, evidence classification, and field verification form an integrated cycle of inquiry. Scientific Recognition Architecture supports the progressive clarification of research objects while keeping explanations open to correction. A proposed pilot using Vietnamese archives provides the setting: a limited documentary corpus would be examined to reconstruct candidate locations and periods for environmental investigation. Archival statements, interpretations, hypotheses, and observations retain distinct evidential status; historical records alone cannot establish present contamination or health effects. AI assists translation, comparison, and the development of alternative explanations, while consequential claims require source checking and specialist assessment. The programme preserves the grounds for decisions, rejected hypotheses, unresolved questions, and revisions across changing projects and participants. Practical online training will prepare an international team to apply these procedures using shared research materials. Participation in native or strongest working languages will be supported through checked translation and common terminology. The pilot will assess whether this organisation improves research questions, error correction, and justified verification priorities.