From the perspective of a scientist working in an international environment, Professor Nguyen Ngoc Diep, Director of the Australia–Viet Nam Strategic Technologies Centre and Director for International Programmes at the University of Technology Sydney (UTS), says the decisive factor for sustainable AI development does not lie in technology itself, but in long-term investment strategies for people, higher education, and the research and innovation ecosystem.

Q: Looking back at the development of AI in Viet Nam over recent years, especially since the adoption of Resolution No. 57-NQ/TW, how do you assess the overall picture of AI development in the country?

A: Viet Nam was among the early countries in recognising the opportunities and potential of AI and therefore began preparing strategies relatively early. Since around 2018, as I recall, Viet Nam organised conferences to promote AI development and application, such as AI4VN. With its strong information technology foundation and a large pool of young engineers, Viet Nam was also quick to master many practical AI applications.

More recently, beyond application, Viet Nam has gradually moved into AI research and development, asserting its position through scientific publications in leading international journals and conferences. For AI researchers, these conferences and journals are like the World Cup in football. Regular participation by Vietnamese research groups, along with the development of AI tools and applications, has helped, and will continue to help Viet Nam demonstrate its potential and turn AI opportunities into real outcomes, not only for Viet Nam but also for the wider world.

Reporter: In your view, what are Viet Nam’s most visible advances in AI development and application today, and in which areas does the country still lag behind its actual potential?

A: There are no shortcuts in science and technology development. As I mentioned, the regular presence of Vietnamese research groups at leading international AI forums clearly shows that Viet Nam is fully capable of developing AI and advanced science and technology domestically. From what I know, companies such as NVIDIA and Qualcomm, after acquiring AI and R&D firms in Viet Nam, have highly rated the innovations developed by Vietnamese scientists working there.

At the same time, universities and research centres in Viet Nam have built strong AI research groups with real-world applications. These are the points I consider to be significant progress in the development of AI and new technologies in Viet Nam: Viet Nam has entered the global “playing field”.

That said, Viet Nam’s potential in AI and emerging technologies is widely recognised by international experts. Viet Nam has a population of over 100 million and a high proportion of young people. Another notable advantage is its solid training in mathematics and basic sciences related to AI and new technologies, which compares favourably with global standards.

This is reflected in the high regard that foreign research institutes, universities, and technology companies have for Vietnamese candidates, alongside those from China and India. With this human capital advantage, Viet Nam could do much more to produce research and applications with global impact. I believe Viet Nam has the potential to develop AI research and applications on the scale of OpenAI or DeepSeek.

Q: If you had to identify the single biggest bottleneck holding back sustainable AI development in Viet Nam today, would it be people, data, institutions, or social trust?

A: To further unlock Viet Nam’s potential to produce globally impactful AI research and applications, the country needs to invest more decisively in its strongest asset: people.

As discussed, the fact that Vietnamese research groups can compete at leading AI forums shows the country is on the right track. But being on the right track is not enough — moving too slowly means falling behind. To accelerate, investment in people is especially important.

At top universities worldwide, the cost of training a researcher typically runs at around 100,000 USD per year. Leading universities often have 1,000 or more doctoral candidates. At UTS, where I work, there are more than 3,000 PhD students. In Viet Nam, even the largest universities may have only around 100 to a few hundred doctoral candidates. This gap is significant if Viet Nam aims to move faster. However, Viet Nam has a cost advantage, as training PhD students domestically is far less expensive than abroad. With stronger investment in young talent, the opportunities are substantial. I hope that one day Viet Nam will have 40–50 universities, each with more than 1,000 doctoral candidates.

To prepare for the goals set for 2035 and 2045, investment in R&D capacity is a critical bottleneck that needs a breakthrough. This means training research engineers through doctoral programmes at Vietnamese universities.

Q: Given this core limitation, what should Viet Nam prioritise to promote innovation while managing the risks associated with AI?

A: All new technologies carry risks, especially AI. However, for Viet Nam to maximise development outcomes, the focus should be on research and application rather than excessive concern over risks. As a later entrant, Viet Nam has the advantage of learning from other countries’ experiences in risk management.

The fundamental constraint in AI development and application in Viet Nam remains investment in people. While Viet Nam has advantages in population size and solid training in mathematics and basic sciences at the high-school level, these strengths have not been fully realised because higher education and research have not yet been developed in line with the country’s potential.

In parallel, building a high-tech start-up ecosystem, including AI, could help Viet Nam take off. Alongside AI, the creative industries present strong opportunities. Leveraging existing strengths and applying AI to creative industries and the digital economy could enable Viet Nam to build companies with global reach. One example is Canva, with a market valuation of around 60 billion USD — significantly higher than that of VinGroup.

Q: Viet Nam has recently enacted its Artificial Intelligence Law. From an international perspective, how do you assess the impact this legal framework may have on AI research, development, and application in the coming period?

A: From my limited experience and personal perspective, I do not have high expectations that the law itself will create major breakthroughs in science and technology development. However, it will help Viet Nam manage AI-related risks.

Real breakthroughs in AI research and development will still depend on investment in people and in the application and start-up ecosystem, within an international environment.

Q: If the recent period is seen as one of building the institutional foundation for AI, which areas should Viet Nam prioritise next to generate spillover effects across the economy and society?

A: It is difficult to predict which specific sectors will deliver breakthroughs or spillover effects through AI. However, if spillover is defined as the creation of Vietnamese technology companies with global influence, the creative industries represent a major opportunity in the AI wave.

With the rapid development and convergence of new technologies such as digital twins, the metaverse, IoT, 5G/6G, and quantum computing, the future of the digital economy and Viet Nam’s potential contribution to the world is huge.

Q: From international experience, particularly in Australia, how important is data for AI development, and what risks does Viet Nam face if data remains fragmented, unclean, or lacking clear ownership?

A: Data is often compared to “oil” in the digital economy, especially for AI tools and AI factories. In Australia and some other countries, data has been prioritised and standardised for quite some time. In the US, for example, the concept of personal credit history has existed for 30–40 years. This determines how much an individual can borrow, and even whether they are eligible for certain jobs. In Australia, regulations require individuals to update their addresses, and companies to report information depending on their size, data type, and incidents, helping to standardise and clean data.

Data may be likened to oil, but to increase its value, it must be refined — much like crude oil. Cleaner, more reliable data is more valuable and reduces processing costs. One current risk is that AI can generate fake data, and distinguishing real from fake data is becoming increasingly difficult with the rise of generative AI.

Since data belongs to users, data exploitation must benefit users, not just the companies that own algorithms or computing resources. This is a challenge facing governments worldwide, not only Viet Nam, particularly in dealing with technology giants such as Meta, Apple, and Google.

Q: You have previously warned about value bias in large AI models. In Viet Nam’s current socio-cultural context, how might this risk affect society if AI is used without appropriate adjustments?

A: Transparency requires knowing what data and algorithms are used to train and refine AI models. However, due to limitations in resources and technology, we often rely on existing AI models without full access to such critical information. As a result, bias and inaccuracies are difficult to avoid or verify.

Like humans, AI can be wrong and imperfect. The challenge, and opportunity, lies in harnessing its capabilities and advantages. Excessive fear of risks can limit opportunities. By developing the capacity to use AI, and through that understanding AI — much like understanding people — we can better manage its risks.

Reporter: Thank you very much for your thoughtful insights. We wish you and your family a happy, peaceful, and prosperous New Year.