ALSA 2025 meeting

Mutual legal assistance (MLA), a treaty mechanism by which one country requires evidence located in another, was built for paper records and now struggles with artificial-intelligence systems whose audit trails are short-lived, cloud-based, and dispersed across jurisdictions. This paper contends that a provenance-over-jurisdiction principle—under which courts admit digital material once its integrity and origin are cryptographically proved, irrespective of physical location—offers a fairer and more workable footing for Asia’s legally plural landscape.
The research design combines two complementary legal methods. First, a PRISMA-guided bibliometric study of English-language sources (2014–2025) maps the distance between scholarship on AI bias and scholarship on transnational evidence. VOSviewer network analysis confirms limited cross-citation, indicating that distributive-justice questions remain undertheorized at their intersection. Second, a close doctrinal comparison examines statutory and soft-law instruments governing digital evidence in Japan, Indonesia, Singapore, and the Republic of Korea, read against the European Union’s risk-tiered AI Act and China’s algorithm-filing rules. Three recurrent pressure points emerge—dual-hash authentication thresholds, minimum log-retention periods, and the availability of rapid-preservation orders—each of which imposes disproportionate costs on parties with limited resources.
Relying on functional-equivalence doctrine and proportionality tests developed by Asian constitutional courts, the paper formulates model MLA clauses that embed the provenance principle while protecting privacy and fair-trial guarantees. The proposal shows how redirecting admissibility toward verifiable origin, rather than territorial seizure, can harmonize divergent regulatory schemes and broaden access to justice for communities harmed by AI-driven decisions across the region, and strengthen trust in cross-border investigative cooperation efforts.