Author: Wang Yilin
Artificial Intelligence (AI) refers to the process and results of simulating human intelligence by computer systems or robots, including perception, understanding, learning, reasoning, decision-making and other abilities. In recent years, with the development of computer technology, Internet platforms, big data analysis, etc., artificial intelligence has been widely used and innovated in various fields, bringing huge changes and value to social and economic development.
Taxation is an important part of national finance and an important means of national governance. The goal of taxation management is to achieve fairness, rationality, efficiency and convenience in taxation, to ensure national fiscal revenue, and to promote economic and social development. With the help of artificial intelligence, big data and other technical means, the tax authorities have accelerated the construction process of the Golden Tax Project. The fourth phase of the Golden Tax Project, which will be launched soon, will focus on promoting the upgrading of tax collection and management methods from "tax collection" to "tax declaration" to "tax calculation". The tax collection and management process has changed from "computer-based" to "online" and then "to the cloud", and the efficiency of tax collection and management has been improved from "experience to manage tax" to "control tax by ticket" to "govern tax by numbers".
The integration of artificial intelligence and tax management uses artificial intelligence technology to collect, analyze, mine and apply tax data, so as to realize the wisdom, digitization and modernization of tax management. The development of AI has gone through five stages. In the fifth stage, AI can realize all aspects of more advanced fields such as machine learning, data mining and artificial intelligence through adaptive learning.
1. The application of AI in the tax industry
The tax industry involves extensive data processing, analysis and reporting, requiring a high degree of professional knowledge and skills. AI can support the tax industry in the following ways:
- Automate data processing. AI can realize the automatic collection, sorting, classification and conversion of data in various formats and sources through natural language processing, image recognition, speech recognition and other technologies, reduce the time and error of manual input and verification, and improve the efficiency of data processing. efficiency and accuracy.
- Intelligent data analysis. AI can realize the intelligent analysis, mining and application of data through machine learning, deep learning, knowledge graph and other technologies, discover the laws, trends, anomalies and values in the data, generate visual reports and suggestions, and assist finance and taxation professionals to carry out decision making and evaluation.
- Innovative data applications. AI can realize the innovative application of data, simulate different scenarios and conditions, predict future results and risks, and explore new solutions and optimization strategies through technologies such as generative confrontation networks, reinforcement learning, and transfer learning.
Based on tax big data "training" algorithm model, so as to identify the "features" of the data and the corresponding results, AI can summarize the historical tax behavior of the enterprise to establish a tax model. After that, AI can make tax predictions through future data predictions and existing tax models.
2. The application of AI in tax management
As the use of analytical tools and technologies is integrated into tax administration, according to OECD statistics, approximately 90% of tax authorities in 52 tax jurisdictions report the use of data science and analytical tools. At the same time, the use of artificial intelligence and machine learning is opening up new avenues for risk management as tax authorities become more proficient at managing large datasets and have more computing power. Tax authorities in more than 70% of tax jurisdictions are already using cutting-edge technology to leverage data to uncover previously hidden assets or identify new risks.
Source: OECD et al (2022), Inventory of Tax Technology Initiatives, https://www.oecd.org/tax/forum-on-tax-administration/tax-technology-tools-and-digital-solutions/, Table DM5
Then in practical applications, AI can assist tax authorities in tax management from the following aspects:
- Through data analysis and risk control, improve the auditing effect of tax authorities and the compliance level of enterprises : using data mining technology, AI can find abnormal or suspicious tax payment behaviors, such as false invoices, from massive financial data, social data, industry data, etc. , tax evasion, profit transfer, etc., and generate risk warnings and audit suggestions for tax authorities. Using deep learning technology, AI can identify and analyze unstructured data such as images, videos, and voices, such as identifying the authenticity of invoices, checking identity information, and analyzing behavioral characteristics, etc., to improve the accuracy and efficiency of tax authorities' inspections. Using reinforcement learning technology, AI can automatically optimize audit strategies and methods according to different audit scenarios and objectives, such as selecting audit objects, determining audit timing, and allocating audit resources, etc., to improve audit effects and cost-effectiveness.
- Carry out knowledge management and intelligent learning to improve the professionalism and innovation capabilities of tax authorities : Using natural language generation technology, AI can automatically generate tax reports, case analysis, policy interpretation and other text content based on tax data and rules. Using natural language understanding technology, AI can perform semantic analysis and understanding of tax texts, extract key information and knowledge points, build knowledge graphs and question-and-answer databases, and support tax officials in knowledge retrieval and learning. Using machine reading comprehension technology, AI can deeply understand and reason complex tax law texts, answer questions from tax officials, and improve tax law application capabilities.
- Provide decision-making support and innovative development, and improve the management level and service capabilities of tax authorities : Using predictive analysis technology, AI can predict future tax revenue and change factors based on historical data and trend analysis, and provide reference for fiscal budgets and policy formulation. Using optimization algorithm technology, you can choose an appropriate optimization algorithm according to different business scenarios and goals, such as linear programming, integer programming, nonlinear programming, dynamic programming, genetic algorithm, etc., to solve optimal tax decision-making problems, such as tax forecasting , tax distribution, tax risk control, etc.
Corresponding to the improvement of tax supervision capabilities, corporate tax management capabilities are also advancing with the times under the blessing of artificial intelligence technology. AI can help enterprises obtain tax knowledge and policy information, and provide intelligent consultation and services. For example, using natural language processing technology, an intelligent question-and-answer system can be built. AI can interact with taxpayers through voice or text, guide enterprises to declare, pay, query and other operations, and answer common questions. Using machine learning technology, AI can recommend suitable preferential tax policies based on the characteristics and needs of enterprises, assist enterprises in tax planning, and plan tax burdens reasonably. Using knowledge graph technology, AI can build a professional knowledge base in the tax field, structure and associate complex tax law rules and cases, and facilitate corporate tax personnel to search and learn. From the perspective of corporate costs, the participation of artificial intelligence in tax decision-making activities can help companies save a lot of operating costs while improving accuracy. At the same time, AI can also monitor potential tax risks in the daily operations of enterprises and promote tax compliance.
The "Guiding Opinions on Accelerating the Construction of a World-Class Financial Management System for Central Enterprises" issued by the State-owned Assets Supervision and Administration Commission of the State Council requires central enterprises to strengthen the five functions of accounting reports, fund management, cost control, tax management, and capital operation, and actively use big data and artificial intelligence. , Mobile Internet, cloud computing, blockchain and other new technologies, and actively explore effective paths to realize financial digital transformation relying on financial sharing. At present, enterprises mainly face challenges such as tax-related compliance management, low work efficiency, and separation of business, finance and taxation. Enterprises need to upgrade tax management and tax platform construction to solve tax compliance, improve work efficiency, group risk control and value creation. .
Based on the world's leading BIP business innovation platform, Yonyou is leading the digitization of taxation and assisting enterprises to build a comprehensive tax management system that is "accurate, compliant and efficient". On the one hand, UFIDA has deepened the business side and the fiscal and taxation side by building a unified taxation platform, and built an integrated digital platform for the industry, finance and taxation system to achieve multi-business data integration. On the other hand, UFIDA makes full use of the elements of tax data assets, strengthens tax risk management and tax analysis capabilities on the basis of data standardization, realizes full life cycle value-added tax forecasting, income tax preferential management, etc., and truly empowers businesses , Optimize the tax burden.
The integration of AI and tax management can not only improve the efficiency and quality of tax collection and management by tax authorities, promote the innovation and improvement of tax policies, but also assist enterprises to efficiently complete tax compliance work and lay a foundation for tax burden planning. But at the same time, AI also brings some new challenges and risks, such as data security, algorithm transparency, and ethical responsibility. Therefore, in the process of promoting the integration of AI and taxation management, it is necessary to fully consider the technical characteristics and social impact of AI, and establish corresponding institutional norms and supervision mechanisms to ensure that AI is safe, controllable, credible and available.