Panel discussion: AI, Data, and the Future of the Civil Service

The roundtable featured international experts, local government representatives, and business representatives. Participants included Dr. Tran Van Tung, Former Vice Minister of Science and Technology, Chairman of the Viet Nam Association for Scientific and Technological Information (VASTI); Dr. Hoang Ngoc Hai, Vice President of APAG; Assoc. Prof. Dr. Tran Quang Dieu, Director, Centre for Technology and Digital Transformation, Ho Chi Minh National Academy of Politics; Mr. Le Thanh Tung, Deputy Director General, Information Technology Centre, Ministry of Home Affairs; Dr. Vincent Wong of the University of Hong Kong; and Dr. Nguyen Dang Quang, Director of the Digital Transformation Program at the University of Finance and Marketing. The discussion was also attended by numerous scholars, researchers, and leaders from ministries, sectors, localities, and universities across 11 countries and territories around the world.

In his opening remarks, Dr. Hoang Vinh Giang emphasized that, following earlier discussions on institutional transformation, implementation capacity, readiness for artificial intelligence (AI), accountability, and inclusive data governance, this session would focus more closely on the day-to-day operations of the civil service, including information processing, decision-making, inter-agency coordination, and public service delivery. The central question of the roundtable was: How can AI and data be used to bring about tangible improvements in the civil service?

Session 1: Changes in civil service operations and practical lessons

Opening the discussion, Dr. Tran Van Tung shared insights from his experience in science and technology management on how AI and data are changing the way public duties are performed. He noted that Viet Nam had taken early and proactive steps to lay the groundwork for AI development, from issuing directives and developing an AI development strategy to the recent passage by the National Assembly of the Law on Artificial Intelligence and related legislation, which establish a legal framework for AI in Viet Nam. He stressed that effective AI development requires high-quality databases and data that are “accurate, complete, clean, up to date, consistent, and interconnected.” In practice, AI has already begun to deliver tangible benefits in administrative procedure processing, support for export-oriented businesses, and legal drafting, helping reduce time and costs while improving work efficiency.

However, Dr. Tran Van Tung also emphasized that AI raises important questions concerning ethics, accountability, and the ability to master technology. Users should employ AI intelligently, verify its outputs, and use it effectively, cautiously, and responsibly rather than relying on AI-generated information without verification.

Drawing on practical implementation experience, Mr. Le Thanh Tung identified three key issues that require particular attention when deploying and using AI.

First, AI must be closely integrated into actual work and serve as a practical solution that helps change the working habits of officials and civil servants. Without a clear connection to specific needs and tasks, it will be difficult to establish AI as part of everyday work.

Second, the ability to formulate effective prompts is crucial. If instructions given to AI are unclear, poorly structured, or misaligned with the intended purpose, the resulting outputs may fall short of expectations and lead users to become disappointed with the technology.

Third, as AI continues to develop in Viet Nam and users remain heavily dependent on international AI platforms, Vietnamese-language data has become particularly important. If such data is not sufficiently rich, comprehensive, and relevant to the needs of Vietnamese users, AI-generated responses may fail to meet expectations, prompting users to return to traditional working methods.

Assoc. Prof. Dr. Tran Quang Dieu noted that, amid digital transformation and the implementation of the two-tier local government model, data has become an important resource. The main challenge, however, lies not in the availability of data but in the ability to extract, analyze, and turn it into actionable information for decision-making.

According to him, AI can help analyze data, identify anomalies, uncover root causes and bottlenecks, and propose more appropriate solutions and operational models. To make effective use of these capabilities, Viet Nam needs to develop Vietnamese-language large language models while localizing and personalizing AI so that it can better understand the country’s language, culture, governance context, and operational characteristics.

One principle he emphasized was that data should be provided and authenticated by the place where it originates, which should serve as its most authoritative official source. Accordingly, grassroots-level authorities should play a key role in generating, managing, and providing data. Specialized units should be responsible for data content and quality, while technology units should provide the platforms, tools, and methodologies.

In addition, tasks should be grouped, processes standardized, and databases developed from the grassroots level. This would enable authorities to maintain control over data storage while ensuring connectivity and data sharing with central-level systems. Digital repositories should be more than storage facilities: they should integrate AI tools for searching, analyzing, and synthesizing information, thereby turning data into an asset for management and administration.

The key message was that, rather than simply asking whether AI should be adopted, Viet Nam should focus on how to master the technology and use it effectively, with a proactive approach to technological ownership.

Offering an international perspective, Dr. Vincent Wong of the University of Hong Kong shared insights into how AI and data can make public services more convenient, targeted, secure, and measurable in terms of results.

He observed that AI is bringing about a transformative shift that requires new thinking and approaches rather than the continued application of traditional methods. Viet Nam should take the lead in adopting AI, he argued, while starting with fundamental issues and carefully weighing the benefits, costs, and risks.

Information security and safety were among the issues he highlighted. For AI systems capable of autonomously performing tasks, access to systems, data, and other resources must be strictly controlled to mitigate potential security risks.

Another concern is cost and return on investment. Deploying AI platforms and software may require substantial financial resources. It is therefore essential to assess the actual value generated by the technology rather than investing simply because AI is new. The key question is: What value will investment in AI create, and how can that value be measured?

In his view, Viet Nam should prioritize applications that can significantly improve the quality of public services. AI could be piloted in public transportation, border checkpoints, and immigration agencies to support traffic and passenger flows and reduce waiting times. It could also be applied in healthcare and specialized fields by drawing on models that have been implemented successfully in other countries.

Viet Nam could also consider piloting AI in areas such as crime detection, facial recognition where appropriate, traffic management, and traffic control. All such applications, however, must be aligned with the country’s practical circumstances and governance requirements.

In particular, as data collection and use expand, legal proportionality and citizens’ rights must be taken into account. Citizens should know why their data is being used and what concrete benefits sharing that data will bring. When the purpose is transparent and the benefits are demonstrated through improvements in people’s lives, public trust and cooperation can be strengthened.

The overarching message was that AI should be adopted proactively but cautiously, with improvements in public service delivery as the key measure of success. Security, safety, investment effectiveness, and citizens’ rights must also be safeguarded. Viet Nam’s notable economic, social, and digital transformation progress provides a strong basis for further testing and developing AI applications suited to domestic conditions.

Dr. Hoang Ngoc Hai emphasized that verification must be a top priority when applying AI, particularly in critical fields. AI-generated outputs should not be assessed solely on the basis of their final results; the process, information sources, and reasoning underlying those results must also be examined.

He identified four areas requiring verification. First, content verification: data must be accurate, complete, clean, and up to date. Second, legal verification: AI-generated outputs must have a sound basis and comply with applicable laws, regulations, and principles. Third, standardization: the format, processes, and methods for using AI-generated outputs must be standardized. Fourth, human verification: users must have the necessary skills to assess information, identify appropriate content, and detect inaccuracies.

AI can support creativity, synthesis, and the processing of large volumes of data, but this does not mean that people can delegate decisions to technology. AI can augment human thinking, but it cannot replace it.

Effective AI users are not those who trust AI unconditionally, but those who know how to control and verify its outputs and take responsibility for the final decisions. In critical areas, humans must remain the ultimate decision-makers and retain final oversight of AI-assisted outputs and decisions.

Dr. Nguyen Dang Quang argued that the development of the digital world is giving organizations a new capability: the ability to observe, diagnose, and forecast on the basis of data.

Whereas organizations previously tended to report events after they had occurred, digital data now enables systems to monitor developments in real time, analyze them, and recommend timely adjustments. This approach is clearly illustrated by ride-hailing platforms, whose algorithms adjust prices based on demand, traffic conditions, and other real-world factors, helping allocate resources according to needs at different times.

Drawing on his experience in university management, Dr. Nguyen Dang Quang noted that a smart organization is not simply one equipped with more technology, cameras, or AI tools. More importantly, it is an organization capable of adapting and allocating resources effectively.

At the University of Finance and Marketing, data can be used to identify course-registration patterns and forecast periods of high demand, enabling the university to arrange classrooms, lecturers, and facilities proactively. Similarly, when processing degree-verification requests, data can help identify peak periods, provide information on processing times, and support demand management, thereby reducing pressure on the system.

In vehicle inspection, a shared data platform could help citizens monitor the status of inspection centers and choose suitable locations and times, rather than crowding into a few centers during peak periods. According to Dr. Nguyen Dang Quang, data mapping combined with AI helps reveal what is happening, while algorithms enable systems to adapt and adjust to real-world conditions. This allows resources to be allocated more effectively, reduces pressure on the system, and improves service quality.

This approach is also important for the future of public governance, as AI and digital platforms can do more than assist with routine tasks: they can help build public services that are flexible, efficient, and centered on citizens’ convenience.

Session 2: Key messages of the panel discussion

During the panel discussion, panelists shared concise messages highlighting the need to proactively adopt, master, and maintain control over artificial intelligence (AI) throughout the digital transformation process.

One key message was “AI First”—treating AI as a new way of working and creating value. AI should be adopted proactively, intelligently, and with appropriate controls, rather than viewed merely as a supplementary technology.

The panelists also noted that AI is creating a new cognitive landscape. While AI will not replace humans outright, people who lack the ability or willingness to use AI may face the risk of being replaced in certain jobs. The ability to use and master AI is therefore becoming increasingly important for both individuals and organizations.

From a governance perspective, AI must remain a tool that serves people; people must not become “tools” of AI or technology companies. AI deployment should be designed to suit the specific characteristics of each country, organization, and sector.

The experts further stressed that AI can enhance and support work but cannot completely replace human beings. People must retain the decisive role in controlling, evaluating, and guiding the use of technology.

Concluding this series of messages, the panelists observed that Viet Nam is entering an era of adaptation, in which the ability to adapt to emerging technologies—especially AI—will be a key factor in the development of organizations and society.

Concluding remarks on the first day of the forum (5 October 2026)

Dr. Bui Phuong Dinh, Vice President of the Academy of Public Administration and Governance, delivered the concluding remarks, summarizing the first day’s proceedings.

The first day concluded after nine presentations and four roundtable discussions. These discussions connected the issue of inclusive governance in a world being reshaped by technology and artificial intelligence (AI) with the practical challenge of turning government capacity into tangible development outcomes.

The discussions showed that public institutions must prepare for change while strengthening their ability to deliver public services in today’s context. Dr. Bui Phuong Dinh therefore highlighted eight key messages.

First, citizens must remain at the center of all reform and development efforts. Sound policies should begin with people’s actual needs, be designed in consultation with citizens and businesses, and be evaluated by the improvements they bring to people’s quality of life.

Second, institutional reform must go hand in hand with implementation capacity. Changes to governance structures will only be effective when responsibilities are clearly defined and sufficient financial resources, personnel, and tools are available to fulfill them. Reform requires continuous attention to how decisions can be translated into action, how shortcomings can be addressed, and who will be held accountable for the consequences.

Third, digital transformation should be leveraged to improve public services. Shared infrastructure, coordinating agencies, and effective handover teams can simplify interactions between citizens and government. Digital investment should be closely aligned with service needs and informed by continuous user feedback. Success should be measured by accessibility, reliability, and the time and effort saved for citizens, while maintaining support for people who have difficulty accessing digital channels.

Fourth, contingency options should become part of everyday policymaking. Governments need to consider plausible future scenarios and assess whether existing policies remain effective as circumstances change. Such options can support planning, spending, investment, and the management of emerging crises.

Fifth, the use and governance of AI should be strengthened. AI has significant potential to improve governance and public service delivery. However, issues of security, fairness, and transparency must be carefully addressed. Public institutions need professional assessment, rigorous human oversight, and clear accountability for every decision. When errors affect citizens, there must be appropriate mechanisms to explain and correct them.

Sixth, investment in people’s skills must continue throughout the transformation process. The capabilities of public servants, educators, and the wider workforce form the foundation of organizational capacity. Education, vocational training, and lifelong learning must respond to economic and technological change. Cooperation among governments, educational institutions, and industry is essential.

Seventh, decentralization and the devolution of authority to communities should be promoted across Asia and, in particular, in Viet Nam. Effective decentralization means more than transferring tasks: it requires granting substantive decision-making authority, accompanied by adequate resources, capacity, and accountability. This is essential to making local governments more responsive and ensuring equitable access to essential services.

Eighth, collaboration should focus on shared outcomes and learning from implementation. Development challenges do not stop at the boundaries of a ministry, sector, or level of government. Data must therefore be shared, resources coordinated, and decisions informed by evidence and community feedback. Ultimately, success should be measured by the quality of services and the well-being of citizens.

Governance innovation is not about building a “modern” government on paper. It is about building a civil service that listens to citizens, has the capacity to act, is prepared to anticipate change, and is accountable for ensuring that the benefits of development reach everyone. To achieve this, institutions, technology, people, and partnerships must move together toward a common goal: turning capacity into outcomes, reform into value, and progress into tangible benefits for citizens.

The second day of the Forum, on 6 October 2026, will take place at the Academy of Public Administration and Governance, 77 Nguyen Chi Thanh Street, Giang Vo Ward, Ha Noi.