Page 207 - ISC PROCEEDINGS 21.4

limitations reflect weak linkages between research institutions, technology firms, and
government agencies, as well as insufficient commercialization of AI technologies. Rather
than broad-based investment, policy efforts should prioritize the creation of sector-
specific AI innovation clusters—for example, in public administration, smart cities, and
digital public services—where government demand can directly stimulate innovation. In
addition, public procurement mechanisms should be redesigned to support domestic AI
firms, thereby creating a demand-driven ecosystem that accelerates the diffusion of AI
technologies within the economy.
Third, improving data governance and interoperability is essential to enable AI
deployment at scale. Although Vietnam has developed substantial digital infrastructure,
data systems remain fragmented and are often managed independently by different
agencies. This institutional fragmentation limits the availability of high-quality datasets
required for AI applications. Therefore, Vietnam should move beyond general
digitalization efforts toward the establishment of a unified national data architecture,
including mandatory interoperability standards, shared data platforms, and clear rules on
data access and usage. Importantly, data governance reform should prioritize cross-
agency integration rather than isolated digital projects, as AI systems depend
fundamentally on large-scale, interconnected datasets.
Fourth, human capital development should be aligned more closely with the
practical needs of AI adoption in the public sector. While existing policies emphasize
digital skills in general, there remains a significant gap in advanced analytical and AI-
related capabilities. To address this issue, training programs should move beyond basic
digital literacy toward specialized AI competencies, particularly for public officials
involved in policy design and data management. In addition, targeted programs should be
developed to bridge the gap between academia and practice, such as joint training
initiatives between universities and government agencies. Addressing human capital
constraints is critical for improving both the Development and Diffusion and Resilience
dimensions of AI readiness.
Finally, a more adaptive regulatory approach is needed to support experimentation
while managing risks associated with AI deployment. Although Vietnam has made
progress in establishing governance frameworks, regulatory mechanisms remain
relatively rigid and do not fully support innovation. Expanding regulatory sandbox
programs for AI applications in public services would allow controlled experimentation
and institutional learning. At the same time, clear and enforceable standards on data
protection, algorithmic transparency, and accountability should be developed to build
public trust—an essential condition for scaling AI adoption.
Overall, the analysis suggests that Vietnam’s main challenge is not the lack of digital
infrastructure, but the limited capacity to convert this foundation into effective AI
capabilities. Addressing this challenge requires a shift from infrastructure expansion
toward institutional strengthening, ecosystem development, and data-driven governance.
By prioritizing policy capacity, innovation diffusion, and data integration—specifically in
response to the weaknesses identified in Table 5—Vietnam can significantly improve its AI
readiness and accelerate its transition toward AI-enabled public governance in the coming
decade.

206