A Collaborative Effort to Advance AI in Vietnam
On March 14, a landmark partnership was forged in Hanoi, signifying a major advancement for artificial intelligence (AI) in Vietnam. Meta, the technology giant, joined forces with the National Innovation Center (NIC), an entity under the Ministry of Finance, to launch the 2025 Vietnam Innovation Challenge. This collaborative endeavor, now in its third year, highlights a sustained commitment to fostering AI development within the nation. The partnership aims to leverage the strengths of each organization to create a robust and impactful AI ecosystem in Vietnam.
The ViGen Project: A Cornerstone of AI Development
The 2025 edition of the Vietnam Innovation Challenge places a spotlight on the ViGen project, an ambitious initiative with far-reaching implications for the future of AI in Vietnam and globally. ViGen is focused on creating a large-scale, high-quality, open-source Vietnamese dataset. This dataset is specifically designed to serve as a crucial resource for training and developing large language models (LLMs). The creation of this dataset is a fundamental step towards building more sophisticated and culturally relevant AI applications.
The core objective of ViGen is to significantly enhance the ability of AI models to comprehend the intricacies of Vietnamese culture, context, and linguistic nuances. By achieving this, the project aims to unlock a wave of groundbreaking AI applications that are specifically tailored to Vietnam’s burgeoning digital economy. These applications are expected to span various sectors, including healthcare, education, finance, and government services. The project’s emphasis on cultural understanding ensures that AI solutions are not only technically proficient but also socially responsible and aligned with Vietnamese values.
Roles and Responsibilities: A Synergistic Partnership
The ViGen project represents a synergy of expertise and resources, with each partner playing a distinct and vital role. This collaborative approach ensures that the project benefits from a diverse range of perspectives and capabilities.
NIC: The National Innovation Center takes the lead in overseeing, coordinating, and ensuring that the project aligns seamlessly with Vietnam’s broader national development strategies. NIC’s role is crucial in providing strategic direction and ensuring that the project contributes to the overall economic and technological advancement of the country. Their involvement guarantees that the project remains focused on national priorities and benefits the Vietnamese people.
AI for Vietnam: This organization, with both technical and financial backing from Meta, is entrusted with the execution of specific components of the initiative. AI for Vietnam brings technical expertise and project management skills to the table, ensuring the efficient and effective implementation of the ViGen project. Their deep understanding of the local AI landscape is invaluable in navigating the challenges and opportunities of developing AI in Vietnam.
Strategic Partners: The project also benefits from the contributions of key strategic partners, including NVIDIA, Viettel, and the Vietnam Academy of Science and Technology. These partners contribute to a vibrant and sustainable cooperative ecosystem. NVIDIA provides cutting-edge AI hardware and software, Viettel brings its extensive telecommunications infrastructure and expertise, and the Vietnam Academy of Science and Technology contributes its research capabilities and academic knowledge. This collaborative network ensures that the project has access to the best resources and expertise available.
Empowering AI with a Deep Understanding of Vietnamese
At its heart, ViGen is driven by a mission to develop a high-quality, open-source Vietnamese dataset that is substantial enough to facilitate the training and evaluation of cutting-edge AI models. This endeavor goes beyond simply enabling AI systems to process the Vietnamese language in a natural way. It also ensures that Vietnam’s ethical standards and cultural values are deeply embedded in the very fabric of AI development. This commitment to ethical and culturally sensitive AI is a defining characteristic of the ViGen project.
The dataset will encompass a wide range of text and data sources, reflecting the diversity of Vietnamese language and culture. This comprehensive approach ensures that the AI models trained on this dataset are capable of understanding and responding to a wide variety of contexts and situations. The open-source nature of the dataset encourages collaboration and innovation, allowing researchers and developers worldwide to contribute to and benefit from the project.
A National Priority: Driving Technological Breakthroughs
Vo Xuan Hoai, deputy director of NIC, emphasized the transformative potential of AI, stating, “AI is transforming the world every day.” He further highlighted the critical importance of the ViGen project for Vietnam: “For Vietnam, developing high-quality, open-source Vietnamese datasets is a key priority to drive technological breakthroughs, innovation, and national digital transformation.” This statement underscores the strategic importance of the ViGen project in Vietnam’s overall development strategy.
The project is seen as a catalyst for innovation across various sectors, enabling Vietnamese businesses and organizations to leverage the power of AI to improve efficiency, productivity, and competitiveness. The focus on national digital transformation reflects Vietnam’s commitment to embracing technology as a driver of economic growth and social progress. The ViGen project is a key component of this broader vision.
Vietnam’s Role in the Global AI Landscape
Professor Yann LeCun, Vice President and Chief AI Scientist at Meta, articulated the broader significance of ViGen and the Vietnam Innovation Challenge. He noted that these initiatives extend beyond mere technological advancements. They serve as a powerful affirmation of Vietnam’s emerging role in the global AI landscape, while simultaneously preserving and promoting the Vietnamese language and culture in the age of AI.
LeCun’s statement highlights the global impact of the ViGen project. By contributing to the development of open-source AI resources, Vietnam is positioning itself as a leader in the field and contributing to the global AI community. The project also demonstrates Vietnam’s commitment to preserving its cultural heritage in the digital age, ensuring that the Vietnamese language and culture are not left behind in the rapid advancement of AI.
“We are not just creating technology,” Yann LeCun emphasized, “we are building an inclusive AI future that stays true to local values.” This statement encapsulates the core philosophy of the ViGen project: to develop AI that is not only technologically advanced but also ethically sound and culturally relevant. This commitment to inclusivity and local values is a key differentiator of the project.
Meta’s Contribution: Open Datasets for Community Benefit
Meta’s commitment to the ViGen project extends to providing open datasets under the AI and Data for Community Benefit program. These datasets encompass a wealth of information, including data on mobility, social connections, and AI-powered population maps. This contribution is poised to propel AI research and applications across a diverse range of fields, from urban planning to disaster response.
The open datasets provided by Meta are a valuable resource for researchers and developers, enabling them to build and test AI models that address real-world challenges. The data on mobility can be used to improve transportation systems, the data on social connections can be used to understand social dynamics, and the AI-powered population maps can be used for urban planning and resource allocation. This contribution demonstrates Meta’s commitment to using AI for social good.
Enhancing Vietnamese Representation in Global AI
Tran Viet Hung, CEO of AI for Vietnam, highlighted the profound impact that ViGen will have on the representation of Vietnamese in global AI datasets. He also pointed out that ViGen will actively contribute to the Open & Trusted Data Initiative (OTDI), a key component of the Global Partnership on AI, in which AI for Vietnam plays a vital role.
The underrepresentation of Vietnamese in global AI datasets is a significant challenge, as it limits the ability of AI models to understand and process the Vietnamese language effectively. ViGen addresses this challenge directly by creating a large-scale, high-quality Vietnamese dataset that will significantly improve the representation of Vietnamese in the global AI landscape. The project’s contribution to the OTDI further demonstrates its commitment to promoting open and trustworthy AI data globally.
Launching the ‘Public Sector Innovation in Asia-Pacific with Open-Source AI’ Handbook
Beyond the ViGen project, Meta and Deloitte have chosen Vietnam as the inaugural country in the Asia-Pacific region to launch a significant handbook titled ‘Public Sector Innovation in Asia-Pacific with Open-Source AI: Unlocking Transformational Potential with Llama.’ This handbook is designed to provide invaluable support to public agencies, enabling them to effectively adopt open-source AI. It serves as a practical guide for implementing AI models that are precisely tailored to local conditions and specific needs.
The handbook provides practical guidance on how to leverage open-source AI models, such as Llama, to improve public services and address local challenges. It covers topics such as data privacy, security, and ethical considerations, providing a comprehensive framework for responsible AI adoption in the public sector. The launch of this handbook in Vietnam demonstrates the country’s commitment to being at the forefront of AI innovation in the region.
Harnessing the Full Potential of AI
Sarim Aziz, Public Policy Director at Meta, underscored the company’s commitment to empowering Vietnamese organizations and businesses: “Through open-source models like Llama, Meta hopes to help Vietnamese organizations and businesses tap into AI’s full potential.” This statement reflects Meta’s broader vision of democratizing AI and making its benefits accessible to everyone.
By providing open-source AI models and resources, Meta is empowering Vietnamese organizations and businesses to develop their own AI solutions and innovate in their respective fields. This approach fosters a more inclusive and collaborative AI ecosystem, where everyone can participate and benefit from the advancements in AI technology.
Real-World Applications: Transforming Government Operations
A report released at the event showcased two compelling examples of how the Llama model has been successfully implemented in Vietnam:
Ministry of Science and Technology: In collaboration with MISA, the ministry has developed a virtual assistant that dramatically reduces the time required for officials to look up information. This has resulted in a remarkable 98% reduction in lookup time, significantly enhancing work efficiency. The virtual assistant uses the Llama model to understand natural language queries and quickly retrieve relevant information from a vast database of documents. This application demonstrates the potential of AI to streamline government operations and improve public service delivery.
Ministry of Justice and Viettel: These entities have jointly applied Llama to create a legal assistant, streamlining the process of document research. This application has led to a 30% reduction in document research time. The legal assistant uses the Llama model to analyze legal documents and identify relevant precedents and regulations, assisting legal professionals in their research and decision-making. This application highlights the potential of AI to improve efficiency and accuracy in the legal field.
These real-world examples demonstrate the tangible benefits of using open-source AI models, such as Llama, to address specific challenges in the public sector. They showcase the potential of AI to transform government operations and improve the lives of citizens.
Open-Source AI: A Driver of Digital Transformation
Chris Lewin, Head of AI and Data Capabilities for Asia-Pacific at Deloitte, emphasized the pivotal role of open-source AI in driving digital transformation within the public sector. He stated, “Through this report, Deloitte aims to help management bodies and organizations in Vietnam gain a deeper understanding of next-generation AI applications based on principles of transparency and trustworthiness.” This statement highlights the importance of open-source AI in promoting transparency and trust in AI systems.
Open-source AI allows for greater scrutiny and collaboration, ensuring that AI systems are developed and deployed in a responsible and ethical manner. The report by Deloitte provides guidance on how to leverage open-source AI to achieve digital transformation while adhering to principles of transparency and trustworthiness. This approach is crucial for building public confidence in AI and ensuring its widespread adoption.
Detailed Explanation of Key Concepts and Initiatives:
Large Language Models (LLMs)
Large Language Models (LLMs) are sophisticated AI systems at the forefront of many AI advancements. They are trained on massive datasets of text and code, enabling them to perform a wide array of tasks, including:
- Text Generation: LLMs can generate human-quality text in various formats, such as articles, poems, scripts, and summaries. They can adapt their writing style and tone to match different contexts and requirements.
- Translation: LLMs can accurately translate languages, capturing nuances and context that traditional machine translation systems often miss.
- Question Answering: LLMs can provide comprehensive and informative answers to a wide range of questions, drawing on their vast knowledge base.
- Summarization: LLMs can condense large amounts of text into concise summaries, extracting the key information and presenting it in a clear and understandable manner.
- Code Generation: LLMs can write code in various programming languages, assisting developers in their work and automating repetitive coding tasks.
The effectiveness of an LLM is heavily dependent on the quality and size of the dataset it is trained on. A larger and more diverse dataset generally leads to a more capable and versatile LLM. This is where the ViGen project’s focus on creating a high-quality, large-scale Vietnamese dataset becomes crucial.
Open-Source AI
The concept of open-source AI is central to the ViGen project and the broader collaboration. Open-source AI refers to AI models, datasets, and tools that are made freely available to the public, typically under licenses that allow for modification and redistribution. This approach offers several significant advantages:
- Transparency: The underlying code and data of open-source AI systems are open for scrutiny, allowing researchers and developers to examine how they work and identify potential biases or flaws.This transparency promotes trust and accountability.
- Collaboration: Open-source AI fosters collaboration among developers and researchers worldwide. Anyone can contribute to the improvement and refinement of the AI models, leading to faster innovation and better performance.
- Innovation: Open access to AI models and datasets encourages a more rapid pace of innovation. Anyone can build upon existing models and datasets, creating new applications and solutions.
- Accessibility: Open-source AI lowers the barriers to entry for organizations and individuals, making AI technology more widely accessible. This democratization of AI empowers more people to participate in the development and application of AI.
- Customization: Users can adapt and modify open-source AI models to meet their specific needs and requirements. This flexibility allows for the creation of tailored AI solutions that address unique challenges.
The Vietnam Innovation Challenge
The Vietnam Innovation Challenge is an annual program designed to:
- Identify and Support Innovation: The program seeks to identify and support innovative solutions to key challenges facing Vietnam, fostering a culture of innovation and entrepreneurship.
- Foster Collaboration: The challenge promotes collaboration and knowledge sharing among stakeholders in the innovation ecosystem, including researchers, developers, businesses, and government agencies.
- Promote Technology Adoption: The program aims to promote the development and adoption of cutting-edge technologies, particularly in the field of AI, to drive economic growth and social progress.
The Significance of Datasets
Datasets are the foundation of AI. They provide the raw material that AI models use to learn and improve. The quality, size, and diversity of a dataset directly impact the performance and capabilities of an AI model.
- Quality: A high-quality dataset is accurate, consistent, and representative of the real-world phenomena it is intended to capture. Errors or biases in the dataset can lead to inaccurate or biased AI models.
- Size: Larger datasets generally lead to better-performing AI models, as they provide more examples for the model to learn from. A larger dataset allows the model to capture a wider range of patterns and relationships in the data.
- Diversity: A diverse dataset includes a wide range of examples, ensuring that the AI model is not biased towards specific groups or perspectives. A diverse dataset helps to create AI models that are fair and equitable.
Cultural and Linguistic Nuances
The ViGen project’s focus on capturing Vietnamese cultural and linguistic nuances is particularly significant. Language is not simply a tool for communication; it is deeply intertwined with culture, context, and identity.
- Cultural Context: AI models need to understand the cultural context in which language is used to accurately interpret meaning and avoid misunderstandings. Cultural norms and values influence how language is used and interpreted.
- Linguistic Nuances: Vietnamese, like any language, has its own unique set of linguistic nuances, including idioms, expressions, and grammatical structures, that AI models must be able to grasp. These nuances can be subtle but crucial for accurate understanding and communication.
By incorporating these nuances into the dataset, ViGen aims to create AI models that are not only fluent in Vietnamese but also culturally sensitive and contextually aware. This ensures that AI applications are relevant and effective for Vietnamese users.
Ethical Standards and Cultural Values
Embedding Vietnam’s ethical standards and cultural values in AI development is a crucial aspect of the ViGen project. This ensures that AI technology is aligned with the nation’s values and priorities, promoting responsible and beneficial AI development.
- Ethical Considerations: AI development raises a range of ethical considerations, including privacy, fairness, accountability, and transparency. These considerations must be addressed to ensure that AI is used for good and does not cause harm.
- Cultural Values: AI systems should reflect and respect the cultural values of the society in which they are deployed. This ensures that AI applications are culturally appropriate and do not clash with local norms and beliefs.
By incorporating these considerations into the dataset and the development process, ViGen aims to promote the responsible and ethical development of AI in Vietnam, ensuring that AI technology benefits all members of society.