Platform interventions arrived late in the cycle and were concentrated in Vietnamese-language content, leaving minority-language communities comparatively unprotected.
This study examines the circulation of election-related misinformation in Vietnam across three major platforms, combining content analysis of flagged posts with interviews with journalists, fact-checkers and platform policy staff.
The research forms part of the Observatory’s work on information integrity, and connects to parallel studies conducted in Sri Lanka, Indonesia and the Philippines. Comparative findings across the four countries are summarised in the regional mapping study.
Methodologically, the study relies on a purposive sample of 4,200 posts collected over an eleven-week period, coded against a shared regional taxonomy so that results can be compared across country studies.
Key findings
- Volume peaked in the final ten days of the campaign, when verification capacity was most constrained.
- Minority-language content was flagged at roughly a third of the rate of Vietnamese-language content.
- Fact-checks reached a fraction of the audience of the claims they addressed, with a median lag of 31 hours.
- Civil-society monitoring groups identified emerging narratives earlier than automated systems in most observed cases.
Figure 1
AI ecosystem readiness score, 2026
Composite of 21 indicators across data, talent, compute, finance and governance. Scale 0–1.
Cite this output
Asia AI4D Observatory (2026). Model cards for low-resource languages: a practitioner toolkit. LIRNEasia, Colombo.