The Information Technology Service Center under the Department of Information Technology, General Department of Taxation

To achieve the goal of reforming the tax system by 2030, the tax sector has expedited the application of information technology to enhance risk identification and clearly identify high-risk businesses that require inclusion in annual inspection plans.

AI applied to risk analysis and management

According to Mr. Vu Chi Hung, Deputy General Director of the General Department of Taxation, the tax industry has increasingly employed AI and big data analysis to regulate electronic invoices. Since electronic invoices are used, the data system has been stored. The system will aid tax officials with tools and data sources to scrutinize, analyze, and detect signs of risk in the issuance and use of electronic invoices. This will prevent and promptly handle invoice violations as well as acts of fraud, tax evasion, and illegal value-added tax refunds. It will enhance the effect of tax management and create a healthy and equal business environment for companies.

With the coordination of related authorities, the General Department of Taxation uncovered 524 businesses selling illegal invoices and falsified invoices. The data on selling/buying invoices is inconsistent. The selling/buying prices of identical goods vary significantly among companies. Some companies only sell but do not buy while others only buy but do not sell. Invoices are often reused to evade taxes. All these unusual signs are easily recognized by tax officials using the big data system and artificial intelligence.

Ms. Nguyen Thi Lan Anh, Director of the Department of SME and Individual Business Tax Management under the General Department of Taxation, stated that the incorporation of big data and AI applications into the database analysis and management system has enabled the easy tracing and detection of suspected illegal purchase and sale of invoices, as well as the prompt identification of illegal use of e-invoices. The tax sector has discovered that some entities have established companies, even a series of companies without any manufacturing/business activity, solely to sell invoices, causing significant losses to the State coffers.

Inspection and examination results are another crucial source of AI. Every year, tax authorities conduct inspections to detect fraud, prevent tax loss, and ensure state budget revenue. In 2021, inspection resulted in VND3,339 billion in tax collection, refunds, and fines, while examination resulted in VND6,791.88 billion. Inspection and examination results can be used as input data to combine unsupervised machine learning training with labeling results obtained through supervised learning algorithms so that the model can continue to identify, verify and classify risks and forecast revenue.

The application of information technology in tax administrative reform has reduced compliance costs for taxpayers and administrative costs for tax authorities. This has minimized interactions between tax authorities and taxpayers, reducing harassment and inconvenience to taxpayers. Additionally, it helps tax authorities, businesses, and banks reduce procedural settlement time; cut costs for businesses and society; and provide public and transparent information for taxpayers.

Gradual development

Despite the progress made, IT applications in the tax industry are still facing challenges. In particular, tax policies, legal documents, and input forms of current documents are constantly changing. According to Decree 102/2009/ND-CP and the Bidding Law, the General Department of Taxation has only 14 months left to develop an IT project that upgrades applications to meet new professional requirements. This has a significant impact on IT applications in the tax industry.

Moreover, the current upgrade, addition, and deployment of IT infrastructure in the tax industry have not met the requirements. The speed of deployment and expansion of IT applications and databases is increasing, which affects the quality and progress of tax performance in the sector. The current IT human resources in the tax industry are still insufficient, and their quality is not guaranteed. Deploying new tax management applications in a centralized manner and according to international standards is very different from previously localized tax management applications. Therefore, there are many difficulties and obstacles that need to be addressed and improved.

AI application to tax risk management is a crucial issue that has been vigorously implemented by the tax sector. However, to enable AI to analyze tax compliance risks more effectively in Vietnam, experts suggest that the tax sector should deploy comprehensive solutions in the near future, such as: developing legal documents and frameworks for risk management using AI; establishing a process for applying risk management in tax administration, a process for collecting and reviewing risk information, and a process for building data infrastructure and information systems; fostering an advanced AI ecosystem to enhance AI applicability through collaboration programs and activities with international organizations and tax authorities of other countries to integrate AI into tax compliance risk analysis, including corporate income tax analysis.

Mr. Mai Xuan Thanh, Acting General Director of the General Department of Taxation, stated that the establishment of the Electronic Invoice Database Center helps tax authorities accelerate big data analysis and AI application to tax management and invoice management, thus facilitating control and prompt detection of invoice violations. The AI application to analyze compliance risks with corporate income tax obligations is an encouraging result and effort of the tax industry to maintain Vietnam’s leadership in tax advantages in Southeast Asia. This is a necessary foundation for the tax sector to complete the Tax System Reform Strategy for 2030.