Bringing AI into public service through real

A recent training conference on the implementation of the Law on Artificial Intelligence, organised by the Ministry of Science and Technology, focused on clarifying new regulations, risk management requirements and ways to put AI into practice in management, administration and the delivery of public services.

AI provides support, humans remain accountable

According to Nguyen Thi Ngoc Trang, Acting Director of the National Institute of Digital Technology and Digital Transformation, AI can help officials search for and compile information, draft and review documents, analyse data, support decision-making and improve the quality of administrative services.

However, the application of AI must also meet requirements for reliability, data security, and the protection of the legitimate rights and interests of organisations and individuals, as well as transparency and accountability. AI development must go hand in hand with AI governance, with people remaining at the centre.

The Law on Artificial Intelligence has established a legal framework for the research, development and application of the technology. To ensure the regulations are effectively implemented in practice, agencies and organisations need to translate them into clear procedures, criteria and responsibilities, and to select the right real-world problems to address, rather than adopting AI simply as a trend.

An important principle emphasised at the conference was that AI should only play a supporting role and must not replace the legal responsibilities of officials and civil servants. AI-generated content and results must be checked and assessed before being used in advisory work or the issuance of administrative decisions.

For systems with potentially significant impact on people’s rights and interests or processing important data, governance requirements must be even stricter.

Such systems need to undergo risk assessments and testing, with their operations logged and monitored throughout their deployment. At the same time, clear responsibility must be assigned to humans for both input data and output results.

Measuring effectiveness rather than counting tools

Drawing on experience in Hai Phong, Nguyen Cao Thang, Director of the municipal Department of Science and Technology, said the city is coordinating the implementation of 18 AI applications.

Each task is assessed by answering three questions: What task will AI address? What indicators will be used to measure its effectiveness? Who is responsible for the input data and output results?

This approach prioritises practical effectiveness over the number of AI tools deployed. The goal is not to introduce as many AI applications as possible into government agencies, but to select the right tasks for AI support, identify the value it delivers and fully control the risks that may arise.

Hai Phong has also determined that AI applications must be accompanied by data standardisation, access controls, information security and training for officials. Solutions need to be piloted and thoroughly evaluated before being considered for wider deployment.

At the conference, several technology companies introduced solutions including virtual assistants, report-generation tools, systems for monitoring the implementation of resolutions, tasks and performance indicators, and tools to support officials in their day-to-day work.

The Ministry of Science and Technology will continue to work with ministries, sectors, localities and businesses in implementing the Law on Artificial Intelligence, while developing infrastructure, data, human resources and the AI ecosystem towards safety, reliability and responsible use.

On December 10, 2025, the National Assembly passed the Law on Artificial Intelligence by a high majority.

The law took effect on March 1, 2026, marking the first time Viet Nam has had a dedicated law on AI and establishing a legal framework for the research, development, application and governance of the technology.