National AI Center: Never hand the mission to those who don't understand the problem

DNHN - Amid the global race for AI that is reshaping technological power dynamics, Vietnam has placed high hopes on establishing a National Center for Research and Development in Artificial Intelligence.

This strategic move reflects the country’s ambition to join the frontline of regional tech leadership.

The Center aims to position Vietnam among Southeast Asia's leading nations in advanced technologies.

Recently, Vietnam officially inaugurated its National Data Center, under the Ministry of Public Security. Acting as the digital economy's heart and lifeblood, this Data Center is expected to accelerate the effective development of the AI R&D Center.

Vietnam’s entry into the AI race presents a tremendous opportunity. Yet, missteps could lead to significant waste of resources. According to seasoned tech entrepreneurs and policymakers, the Center's success will not depend on the scale of investment or the number of deployed applications, but on strategic thinking.

In an interview with Business & Integration Magazine, Mr. Le Viet Thang, Founder and CEO of 1Office, the AI Agent platform powering over 6,000 businesses shared his views. With over a decade of real-world experience, he emphasized that building a National AI R&D Center must be a long-term investment in effectiveness, not a performative endeavor.

The pain points hindering AI development in Vietnam

Currently, Vietnam faces a major bottleneck: data. Following administrative boundary mergers, new units emerged, but data systems remain fragmented and incompatible. This fragmented landscape creates one of the biggest obstacles for AI: disjointed, non-standardized, unconnected data. Vietnam’s data infrastructure resembles a patchwork quilt, lacking cohesion and standards, a serious barrier to training AI models.

From Mr. Thang’s perspective, the journalist realized four key pain points are choking Vietnam’s AI potential:

Pain Point 1: Poor-quality, inconsistent data. There is no nationwide data standard. Businesses lack clean, sufficient data to build models. However, with the National Data Center now operational, this issue is expected to be addressed more effectively, especially with the Ministry of Public Security at the helm.

Pain Point 2: Lack of sandbox testing environments. Legal and procedural barriers deter businesses from piloting AI in fields like education, agriculture, and public administration. Without regulatory sandbox models, companies are afraid to make mistakes. And without the freedom to err, there can be no innovation.

Pain Point 3: Tokenistic AI projects. Many initiatives use AI as a buzzword. Chatbots and virtual automation are deployed without real users or clear problems. No one wants to invest in unproven technology.

Pain Point 4: SMEs are left behind. Most Vietnamese businesses are small or micro-sized. While agile, they lack tools. Today’s AI is either expensive, complex, or irrelevant. This massive market remains untapped. "Vietnamese SMEs are incredibly adaptive. The problem is that the solutions haven't arrived yet," said Mr. Thang.

Lessons from successful global AI Centers

International experience reveals that success lies not in size or architecture, but in ecosystem operations. Key drivers include government leadership, enterprise initiative, robust data infrastructure, ethical standards, and controlled experimentation.

To be truly impactful, the Center must function as an ecosystem, a hub for data, talent, regulatory frameworks, and markets, not just a symbolic institution. The real goal is to make it the nerve center of Vietnam’s AI ecosystem.

What businesses say: Don't be symbolic, be strategic

Mr. Thang stressed, "Vietnam doesn't need a showpiece AI center. We need a sustainable innovation core where government, business, and academia co-exist to foster responsible, practical, and human-centered AI."

"The government shouldn’t do it all. Let businesses lead. The government just needs to do its part, and that alone is a win."

Only the government can consolidate national datasets. No enterprise can achieve this alone. Therefore, the National AI Center should not focus on operations or production, but on building open, controlled datasets, enabling testbeds, issuing ethical standards, and fostering a vibrant innovation network.

"When it comes to practical application, let the businesses handle it. They understand the problems and are accountable for results. The government has issued many mandates, but none have been delivered because those defining the problems don't understand them."

In the AI race, speed matters. But only the right direction ensures victory. Past failures stemmed from misaligned objectives and a lack of business involvement.

Blueprints for real-world impact

National AI Centers only matter when they avoid performative strategies and focus on solving real national issues. Top global AI centers didn’t aim to show off super apps or replace human creativity. They aimed to enhance productivity and human potential.

Institutes like MIT CSAIL (USA), Stanford HAI, and AI Singapore succeeded not due to state budgets but due to close ties with the private sector and purpose-driven missions. MIT CSAIL thrives through links with Amazon, NVIDIA, and IBM. Stanford HAI pioneers ethical AI, and AI Singapore, though nonprofit, adopts a results-oriented enterprise model with measurable annual impact.

Key principles from global experience

• Government as enabler: Open data, adaptive policies, seed funding—no micromanagement.
• Businesses as executors: They understand the market, innovate fast, and are accountable.
• Focus on pain points: Address domestic issues like public services, healthcare, education, and local language processing.

Success comes not from doing everything but from doing the right things, at the right time, with the right capabilities.

If Vietnam’s National AI Center defines its core problems clearly, such as AI for public management, local dialect translation, district-level healthcare analytics, or unstructured document processing, its chances of success will soar.

Expert recommendations for Vietnam’s AI Center

1. Build a controlled, standardized national open data repository.
2. Establish AI sandbox environments in ministries and localities.
3. Prioritize local AI solutions in public systems as practical support for Vietnamese tech firms.
4. Form an independent Ethical & Standards AI Council aligned with Vietnam's culture and human values.

"The Center's first move must be to establish a dynamic, standardized, and expandable national open data system," Mr. Thang emphasized.

To be a true innovation hub, the National AI Center must break away from formality and become a problem-solving engine, not just a technological label. 

Hoang Lan