Harnessing AI to Support Public Sector

Artificial Intelligence (AI) is seen as a new core driving force that fosters innovation, growth and sustainable development in the world. In Vietnam, the public sector plays an important role in shaping and leading the AI development landscape by integrating AI into governance and operations, delivering public services to people and promoting the domestic AI ecosystem.

Adopting technology is expected to significantly streamline administrative procedures

New progress in the public sector

According to a recent report released by the United Nations Development Program (UNDP) in Vietnam and the Institute for Policy Studies and Media Development (IPS), several central and local government agencies in Vietnam have begun applying AI in public administration and service delivery.

Notable examples include virtual assistants in public administration, facial recognition software for public security, and intelligent traffic monitoring systems. One of the typical successful cases is the use of virtual assistants for judges. The application started to be built in 2022, before ChatGPT, based on traditional AI models. Virtual assistants support judges with a number of tasks such as finding judgments and synthesizing documents and are now widely used in courts at all levels and considered to be effective, able to help reduce the workload by 30%. According to statistics, by mid-2024, there were 173,206 legal documents, 27,610 questions and answers to legal situations and 1.4 million judgments made by judges using integrated virtual assistants. Locally, Tay Ninh province is one of the typical examples of AI adoption efforts like developing and integrating AI into surveillance cameras and applying AI to assist civil servants in handling feedback and recommendations.

Mr. Patrick Haverman, UNDP Deputy Resident Representative in Vietnam, said: AI is transforming how we live, work and govern. It offers remarkable opportunities - but only if it is used responsibly. Vietnam is at an important time; with the right strategy, investments, and ethics in place, AI can be a powerful accelerator of national development and human progress.

In fact, AI application in the public sector helps save resources and time while enhancing work performance of both public employees and agencies. However, AI application in the public sector has recently revealed several weaknesses like low practicability, insufficient legality, inadequate guidance and lack of attention to AI risks.

In particular, the role of data and infrastructure in effective AI application is critical. Currently, limited data sharing and connectivity between government agencies hinder large-scale AI solutions. Therefore, policies promoting open data, simplifying administrative procedures related to data, and investing in computing infrastructure and data centers for AI are essential.

Moreover, limited human resources and technological capacity are significant challenges.  According to Dr. Nguyen Xuan Hoai, Director of the Institute of Artificial Intelligence (University of Technology, Vietnam National University, Hanoi), currently, there is a critical shortage of AI personnel. Each year, only 10% of recruitment needs are met while only about 30% of 55,000 IT students graduating each year can work in AI.

In addition to the shortage of AI human resources, Vietnam faces other challenges such as lack access to leading AI experts and consultants to evaluate products needed by the market and lack access to infrastructure, platforms and tools ready for businesses. The number of AI experts in the public sector is small, leading to heavy reliance on the private sector.

In addition, unclear policies and mechanisms pose major obstacles. Although Vietnam has an AI strategy, the legal framework and technical standards are incomplete, especially in data management and AI ethics. Financial challenges are also notable, as AI implementation requires substantial resources, but the state budget is limited, and financial mechanisms are inadequate, making investment in AI applications difficult.

Choosing right AI technology

Mr. Nguyen Quang Dong, Director of IPS, said that AI application in Vietnam's public sector holds great potential, particularly in enhancing policymaking capabilities, improving work efficiency, and advancing public service delivery. Therefore, each agency must unlock this potential by clearly identifying its specific challenges to select the appropriate AI technology while also assessing its internal readiness to develop an effective implementation roadmap, he added. Government agencies must enhance the skills of civil servants, leverage private-sector expertise, and foster collaboration between stakeholders in AI development.

Besides, according to experts, Vietnam needs to enhance digital technology capacity in government agencies by investing in data, data infrastructure and high-performance computing systems for AI while strengthening human resource training for AI development and operation. It is necessary to establish a clear legal framework and mechanisms, including regulations on accountability, ethical standards, risk management and data privacy protection.

Notably, it is necessary to foster public-private partnerships to encourage technology companies and research institutions to develop AI products and solutions for the public sector; promote open data and data governance and create a highly connected system among government agencies and partners; and focus on ethical AI and risk management through monitoring systems, impact assessments, and ensuring transparency, fairness and privacy protection.

However, experts also noted that Vietnam needs to apply AI selectively, practically and appropriately to actual conditions as well as specific needs in the public sector, and avoid simply chasing after technology trends, to ensure real efficiency and public benefits.

Experts have proposed two strategic options for AI development in Vietnam. The first is narrow, sector-specific AI: Cost-effective, leveraging existing data, easy to control, and delivering immediate results. The other is basic AI research and large model development (like LLM): Deep research to master core technology but require substantial financial resources and advanced computing infrastructure.

The implementation roadmap will need to correctly identify the right problem: Clarifying needs and prioritizing AI applications that fit each agency's reality; experiment and refine: Testing AI solutions in a controlled environment and adjusting based on real-world feedback; and scale up and collaborate: Expanding implementation after proving effectiveness, establishing a legal framework, and promoting public-private and international partnerships to enhance AI resources and expertise.

By Anh Mai, Vietnam Business Forum