Page 103 - ISC PROCEEDINGS 21.4

change. At the firm level, CUDE Journal (2023) confirms positive but heterogeneous
digital technology effects on TFP, varying significantly by firm size, sector, and initial
capability.
On AI adoption, the NIC-JICA-BCG Vietnam AI Economy 2025 report projects AI
could contribute up to $130 billion, roughly 25% of Vietnam's current GDP, by 2040.
Google's (2024) policy research tempers this optimism by identifying skills gaps,
infrastructure barriers, and data governance deficiencies as the primary constraints. ERIA
(2023) underscores the systemic nature of these challenges, noting that firms require
support across all stages of digital transformation. The OECD Economic Surveys: Viet Nam
2025 similarly calls for scaling R&D investment, strengthening university-industry-
government linkages, and expanding PhD-level research capacity.
The critical research gap this paper addresses is the absence of a unified framework
integrating: (a) Vietnam's evolving policy and legislative architecture of 2024-2026; (b) the
productivity paradox literature; and (c) sectoral AI deployment evidence across
agriculture, manufacturing, healthcare, and public administration, a synthesis that prior
literature has not achieved.
3. Research methodology
This paper employs a qualitative, document-based methodology grounded in
systematic secondary-source synthesis, appropriate for research concerned with
diagnosing structural conditions and formulating policy recommendations, which require
analytical depth and cross-contextual interpretation rather than primary quantitative
inference. The approach integrates four complementary techniques. First, systematic
literature review drawing on academic journals, international organization reports (World
Bank, OECD, McKinsey Global Institute, UNDP), and peer-reviewed publications in
economics, management science, and information systems, assessed for credibility,
recency (prioritizing 2020-2026), and relevance. Second, comparative analysis
benchmarking Vietnam's AI-productivity landscape against South Korea, Singapore, and
Thailand to identify transferable lessons. Third, policy document analysis covering key
Vietnamese legislative and strategic instruments, including Politburo Resolution No. 57-
NQ/TW (2024), the National AI Strategy (2021, updated 2025), the AI Law, Decision No.
749/QD-TTg (2020), Decision No. 411/QD-TTg (2022), the 2026 socio-economic
development plan, and Resolution No. 198/2025/QH14. Fourth, combined deductive and
inductive reasoning, applying GPT theory, the productivity paradox, and TFP
decomposition to generate diagnostic findings, while inductively deriving policy lessons
from sector-level evidence.
Primary data sources include official statistics from Vietnam's GSO and NSO, the
World Bank, OECD Economic Surveys: Viet Nam 2025, and the e-Conomy SEA 2025 report,
supplemented by the Vietnam AI Economy 2025 report, Google's AI Opportunity Agenda
(2024), CSIRO Aus4Innovation reports, and survey data from Microsoft, Cisco, and
industry associations. The principal limitation is reliance on secondary data, which
precludes causal inference at the firm or sector level. Panel econometric studies on
Vietnamese firm-level data would strengthen future causal estimates, a limitation
explicitly acknowledged in the conclusion.
4. Analysis of the current state of AI and productivity in Vietnam's digital economy
(2021-2025)
4.1. Digital economy expansion and AI deployment: achievements and sectoral
progress

102