Expanding AI applications in state agencies

First of all, at the orientation level, from some scattered contents in the National Strategy on AI, there needs to be a general, more clearly defined general content on AI application in the entire State sector from the Central to the local level, from goals, general principles, to key tasks and roadmap. Developing AI in state agencies aims to improve the efficiency and quality of operations of state agencies; both serving the people better, aiming at the interests of the people, in accordance with human ethical standards such as safety, accountability, transparency and inclusion.

At the same time, the legal framework on issues related to AI application in state agencies is very important. This framework includes contents such as: rights and obligations of state agencies in applying AI to their operations, scope of AI application; principles of transparency and accountability; standards for purchasing, designing, developing and using AI-based systems; guidance process on professional and technical aspects of AI application; identification of risks and risk management in AI application; things that cannot be done in applying AI to state agency operations.

Next, correctly identifying and solving the “problem” to deploy AI in state agencies is of practical significance, based on the following factors: (1) determining the need to apply AI according to the specific tasks of state agencies; (2) determining the potential and risks of AI; (3) the capacity of state agencies in terms of infrastructure, data, human resources, and financial resources. In the short term, given that financial, data, infrastructure, and human resources are limited, a targeted, “cut your coat according to your cloth” approach is appropriate. In the medium and long term, when resources are increased, state agencies will need to redefine the “problem” and can develop larger-scale, more complex AI solutions.

In terms of finance, in order to be able to calculate and spend money on AI solutions, which have many unique features compared to information technology, there needs to be specific and separate legal regulations on investment management for AI from the state budget, from budgeting, management, use, and settlement of funds; pricing methods and norms for expenditure items such as data collection, synthesis, testing, and cleaning; training and refining AI models/applications; operating AI solutions in work, etc. In particular, there needs to be a financial mechanism that accepts "trial and error" to a certain extent in applying AI to the operations of state agencies.

Regarding data and infrastructure, connecting and sharing public sector data, providing better open data to businesses and research sectors are indispensable conditions for developing AI solutions for state agencies. In the short term, due to limitations in data and AI infrastructure (data centers, AI chips), it is necessary to develop simpler AI solutions that use less computing power such as virtual assistants for officials and civil servants; document review. In the medium and long term, to upgrade current AI solutions, or develop complex AI models/solutions on the environment and agriculture, it is necessary to invest more in big data, cloud storage infrastructure, computing infrastructure, and AI chips. State agencies can use data center infrastructure services, cloud computing platform services, and data storage services for AI from businesses.

Finally, to have human resources capable of applying AI in the public sector in Vietnam, for the core group of human resources in charge of AI application, it is necessary to pay attention to improving knowledge, professional skills, and techniques in AI, data, cloud technology, network security, etc. It is possible to provide funding for this group of human resources to participate in long-term and short-term training courses on the above contents. For other sectors, it is necessary to train and foster AI, how to use AI in specialized fields; integrate knowledge about AI in current training and fostering programs for officials and civil servants.

Nguyen Duc Lam