Autonomy and identity in AI development

On a global scale, the US, China, and Europe are the top centers in the AI race, each with their own unique approaches. For Vietnam, this race highlights the importance of national control over AI development. Investing in AI research, high-tech talent, and creating “Made in Vietnam” AI models would strengthen the country’s competitiveness while safeguarding data sovereignty and national interests.

Powered by language

According to Mr. Andre-Louis Rochet, Managing Director and Co-founder of TW3 Partners from France, surveys reveal that, despite being a new technology, AI is receiving the highest level of consensus from leaders in Vietnam. “This is important, because a high level of consensus will determine the growth speed and level of implementation later on,” he said. “Vietnam is a fast-growing economy, with a culture that is not afraid to experiment with new things and implement big plans. This spirit is also reflected in how the country approaches technologies like AI.”

Notably, Mr. Rochet pointed out that countries around the world are developing their own AI strategies, with each market having its own “Big Tech” players. For example, the US has Microsoft, OpenAI, and Anthropic, Europe has Mistral, and China has made rapid progress, catching up with the US with key companies like Alibaba, Tencent, and DeepSeek, the latter of which has developed models that even surpass OpenAI’s leading models. In this context, countries like Vietnam can choose to rely on US models, use Chinese models, or adopt European models. However, Mr. Rochet believes these models may not be “suitable” for Vietnam.

The major challenge, he continued, is that Vietnam needs to develop its own AI models, using Vietnamese-language data that reflects its language and culture. Without a swift response, as most data generated today is produced by AI and primarily in English, Vietnam’s AI models may be overwhelmed by English-based models, making it increasingly difficult to train Vietnamese-language models.

Dr. Tran Viet Hung, CEO and Co-founder of the AI for Vietnam Foundation, believes that the potential of AI in Vietnam has yet to be fully tapped. He argues that with the rapid development of AI, if Vietnam does not seize the opportunity in time, it might miss out, and “one of the most important factors is supporting Vietnamese in AI,” he said.

In datasets used to train AI models, he continued, Vietnamese data accounts for only about 0.3 to 0.8 per cent. As a result, when using current AI models, though they may provide valuable information, the results are often not natural and do not accurately reflect Vietnamese culture. This leads to AI applications in Vietnam not being fully effective and not realizing their full potential. Dr. Hung believes that to leverage and develop the power of AI in Vietnam, supporting the Vietnamese language is crucial.

Balancing innovation and independence

In an interview with Vietnam Economic Times / VnEconomy, Professor Tran Thanh Long from the Department of Computer Science at the University of Warwick in the UK emphasized that, in the “AI feast”, if a company wins, they will gain complete control. For instance, when OpenAI or Microsoft sells products to countries and businesses, even though they are in the “customer” role, these countries and companies will become dependent on the systems of these companies, and any change or adjustment, no matter how small, must go through them. “This leads to many consequences, not just financially, but also in terms of national security,” Professor Long said.

AI systems are evolving at an incredibly fast pace, he went on. Decisions that seem beneficial now may become a threat within just a few years or even months. This is the essence of “technological colonization”; dependence not only in terms of finance but also in security, data, and core technology control. When this dependence becomes too significant, even minor changes in the system will be subject to external forces, causing a loss of sovereignty.

Therefore, Professor Long suggested that Vietnam needs to clearly define which areas should remain “off-limits”, which can be used for collaboration to learn and adopt technology, and which can be fully open to foreign investment. First, certain sectors should be strictly protected from foreign intervention, especially those related to national security, defense, and the police.

Second, there are areas where collaboration can take place to learn from others, but the focus should be on truly acquiring and transferring technology and knowledge. Vietnam needs a strategy to deeply integrate technology, even learning how to seize opportunities to access and master knowledge from foreign partners.

Third, some sectors can be open to foreign investment, allowing participation and profit-making. However, even in these areas, Vietnam needs a tight management strategy to ensure national interests and avoid total dependency.

Apart from security and defense, Vietnam should also protect two other key sectors: healthcare and education. These fields directly impact the country’s long-term development, and thus need strict control to avoid excessive foreign influence.

He suggested that to resolve this issue, the government must take the initiative in investing and making decisions; it cannot just wait for businesses to act. Right now, even large Vietnamese enterprises still have to buy machinery, rent servers, and store data on foreign platforms like Amazon Web Services or Google Cloud. “If the government doesn’t have clear policies, businesses will continue to rely on foreign entities,” Professor Long said.

He also noted that Vietnam needs a clear strategy, where servers and important data infrastructure should be located within Vietnam and have strict controls over data input and output. Furthermore, Vietnam could learn from Europe in creating a “playing field” for domestic businesses and individuals to develop technology.

One positive factor is that the world now recognizes the importance of diversifying the technology supply chain, and many foreign companies are willing to collaborate with Vietnam to develop these systems. This is an opportunity that Vietnam can seize to build autonomous technology infrastructure in the future.

Mr. Rochet also believes that Vietnam’s AI future will be defined by digital technology autonomy and cultural identity. Developing AI that is uniquely Vietnamese, trained in Vietnamese, and aligned with the data and needs of Vietnam will help reduce dependence on foreign technology while preserving the nation’s cultural identity in the process of technological development.

According to Mr. Rochet, “uniquely Vietnamese” AI goes beyond just speaking Vietnamese. First, it is trained using Vietnamese data. This AI is developed based on domestic data, reflecting the language, culture, and specific needs of the Vietnamese people. Instead of using English or Chinese data, this AI will naturally understand the context, idioms, and communication styles of the Vietnamese.

Second, it aligns with cultural identity. Vietnamese AI will not only be a technical tool but also a means of preserving and promoting national identity, creating content that fits the traditions, history, or lifestyle of the Vietnamese, rather than being influenced by foreign cultures.

Third, it ensures technological autonomy. Developing uniquely Vietnamese AI means Vietnam will own core technologies, reducing dependence on foreign platforms. This will help protect data sovereignty and increase competitiveness in the international arena.