INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1619
METHODS OF USING ARTIFICIAL INTELLIGENCE IN THE ANALYTICAL
MECHANISMS FOR DETERMINING THE RISKS OF LAW VIOLATIONS IN THE
FIELD OF STATE EXPENDITURES AND REVENUES
Abdullaev Mehriddin Razzoqovich
Doctor of Phylosophy in technical Sciences(PhD)
The Accounts Chamber of the Republic of Uzbekistan
Fayziev Shavkat Ismatovich
doctor of technical sciences (DSc),docent
The Account Chamber of the Republic of Uzbekistan
Atoev Asliddin Ekhtiyorovich
The Accounts Chamber of the Republic of Uzbekistan
Abstract:
The article describes the role of artificial intelligence technologies in identifying
the risks of violations of the law in the field of public expenditure and revenue, and details
the definition, functions, differences and interrelationships of artificial intelligence, machine
learning and neural networks.
Key words:
budget funds, risks, information technologies, artificial intelligence, machine
learning, neural networks.
Introduction.
Artificial intelligence has become one of the main tools of modern society and
has a significant impact on various areas of human life, including public administration in
developed countries. With the help of artificial intelligence, it is possible to perform financial
forecasting of budget income and expenses and to increase their effectiveness, to reduce the
risk of making mistakes and errors, to carry out risk analyzes without the human factor, and
to continuously monitor budget expenses and income.
It is worth emphasizing that artificial intelligence is able to quickly adapt to innovations and
changes introduced in legislation and by-laws and can become an important element in
creating a transparent and stable state financial control mechanism.
Let's look at a few examples:
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1620
1. Using artificial intelligence to forecast budget expenditures:
- expenditures of previous years;
- Presidential resolutions and decrees;
- makes it possible to forecast the expenses that should be carried out in connection with state
programs and identify inefficient expenses through training.
2. Budget revenue forecasting uses artificial intelligence to estimate budget and non-
budgetary revenue based on various algorithms and scenarios, which helps in more accurate
budget revenue forecasting.
3. Monitoring anomalies in budget fund transactions, i.e. artificial intelligence analyzes
financial transactions and helps identify suspicious or anomalous operations. This makes it
possible to detect cases of embezzlement of budget funds.
4. In automated auditing, artificial intelligence automatically analyzes accounting data and
reports to identify errors or irregularities. This allows errors to be detected without
inspections.
5. With the help of artificial intelligence, the data required for audit activities is automatically
collected. Based on this information, errors and omissions and violations of the law are
automatically detected, as well as reducing the duration of audit activities.
6. By using artificial intelligence to analyze large amounts of data, internal audit staff will be
able to monitor the required statistical and other indicators online and identify any
uncertainties.
7. Artificial intelligence makes it possible to assess risk levels by taking into account many
factors such as external and internal economic conditions, government project costs,
deadlines and past mistakes.
8. Economic trends and investment opportunities are analyzed through the use of artificial
intelligence in the management of state assets and liabilities, which helps in effective
management of state assets and liability forecasting, as well as in modeling the expected
future state (debt, pension, etc.) based on historical data.
9. By processing contracts related to budget expenditures, it analyzes risks using artificial
intelligence, helping to analyze potential violations of contract terms or deadlines, and the
amount.
It is clear that artificial intelligence is a powerful tool for increasing the efficiency of state
financial control bodies and internal audit structures, helping to process large amounts of data,
identify risks, and optimize processes.
Today, artificial intelligence, machine learning, and neural networks are widely used to
process large amounts of data in unconventional ways.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1621
The differences between neural networks, machine learning, and artificial intelligence are
reflected in their levels of abstraction, functions, and technologies. All of them are
interconnected and are used in different ways in the development of intelligent systems.
Below is a definition of artificial intelligence, machine learning and neural networks, their
functions, examples, differences and interrelationships:
1.
Artificial Intelligence (AI)
Definition: This is a broad field that focuses on creating machines or programs that perform
tasks that require intelligence, such as understanding language, recognizing objects, making
decisions, and solving problems.
Function: Used to create systems that mimic the human mind, such as logical thinking,
planning, and data collection or processing.
Examples: Virtual assistants, autonomous vehicle control systems, or chess-playing programs.
2.
Machine learning (ML)
Definition: Machine learning is a specific area of AI that focuses on creating algorithms
that “learn” from data. Instead of programmed rules, ML systems use statistical methods to
improve their decisions as new data comes in.
Task: machine learning is used to analyze data, classify it, make predictions, and automate
the decision-making process.
It is divided into several types:
- teaching with a teacher;
- teaching without a teacher;
- strengthening knowledge through re-learning.
Examples: recommendation systems (YouTube, Netflix, etc.), fraud detection algorithms,
market demand forecasting, etc.
3.
Neural networks
Definition: Neural networks are a type of machine learning technology. Inspired by the
structure and function of biological brain neurons, they consist of artificial neurons arranged
in multiple layers. They are trained on large amounts of data and can solve complex tasks
such as image recognition or voice recognition.
Purpose: Neural networks are particularly effective for tasks involving large and complex
data sets, where traditional machine learning algorithms are not very accurate. Deep neural
networks (deep learning), which use a multi-layer architecture, allow for efficient and fast
solutions to tasks related to processing images, sounds, and text.
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1622
Examples: Facial recognition systems, voice assistants like Google Assistant, processing of
different languages (e.g. translations or chatbots).
The main differences:
Category
Artificial Intelligence
(AI)
Machine learning
( ML )
Neural networks
Definition
Intelligent systems of
creation wide field
SI's
information
based on taught
part
Complicated
tasks
solution to do for of
neurons many multi-story
modeling method
Purpose
Human mind imitation
to do
Decisions reception
to do for in the data
teaching
Many multi-story neurons
through
complicated
patterns modeling
Information -
from useful -
what
From the data use
possible
,
but
mandatory not
Big
in
volume
information
demand will be
done
Big in volume information
in deep trouble used
Example
Virtual assistants ,
autonomous
vehicles
management system
Recommendation
to do systems ,
forecasting
Face familiar , voice
orders understanding
Dependencies:
Artificial intelligence is a general concept of creating intelligent systems.
Machine learning is one of the methods used to implement artificial intelligence, where
systems are trained on data.
A neural network is a machine learning technique used to work with big data.
These concepts are related, but they differ in their tasks, methods, and applications.
Let's consider these technologies in the context of detecting violations in public procurement.
According to Article 46 of the Law of the Republic of Uzbekistan "On Public Procurement",
it is prohibited to illegally select non-competitive methods of public procurement during the
public procurement process, to influence public procurement entities, Disclosure of
information about the participation of participants in public procurement, unreasonably
limiting their number or increasing the requirements for their qualifications, and other forms
of preventing, restricting, or eliminating competition are not permitted.
Article 21 of the Law of the Republic of Uzbekistan "On Competition" prohibits unfair
competition by business entities or other persons acting in their interests against competitors,
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1623
It is prohibited to take actions that would hinder the entry of a competing economic entity
into the commodity or financial market or lead to its exclusion from the market.
For example, a budget customer who had previously colluded with a criminal organization
concluded a contract No. 1 for the purchase of "ryazhenka" (product code 123456789)
through an electronic store with LLC "Product Supplier" for 1.1 million soums (the product
characteristics, which were not available for selection in the electronic system and were filled
at the discretion of the supplier, were indicated as yogurt).
Based on invoice No. 1 under this contract, 100 percent payment was transferred by the
customer to the supplier company.
Option 1:
When the goods were received on the spot, the supplier delivered the "yogurt" product in the
invoice instead of the "ryazhenka" product in the contract, and the customer accepted this
product.
The following proposals have been developed to prevent this situation:
1.
using artificial intelligence to compare the product names entered by suppliers in
electronic procurement systems with the established product codes, identify discrepancies
and implement a mandatory correction mechanism by the system;
2.
Identify discrepancies in subsequent control measures by comparing the products
specified in the contract with the products specified in the invoice using artificial intelligence.
The algorithm of the practical application of the above suggestions is given below:
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1624
Conclusion. In order to effectively use the possibilities of artificial intelligence in the
planning of the state budget, in the process of state projects and procurement, the following is
suggested:
Consolidation of information available in ministries and agencies into a single database,
bringing it to a systematic structured view;
Identify cases of violations of the law by performing risk analysis of the collected data with
the help of artificial intelligence;
Sending to relevant ministries and agencies and law enforcement agencies in order to
eliminate identified risks.
At the same time, the above proposals and mechanisms are reflected in Appendix 3 to the
Decree of the President of the Republic of Uzbekistan No. PF-100 dated July 10, 2024 “On
INTERNATIONAL JOURNAL OF ARTIFICIAL INTELLIGENCE
ISSN: 2692-5206, Impact Factor: 12,23
American Academic publishers, volume 05, issue 04,2025
Journal:
https://www.academicpublishers.org/journals/index.php/ijai
page 1625
additional measures to strengthen financial control over the use of budget funds”, The
Accounts Chamber of the Republic of Uzbekistan has approved a list of 84 ministries and
departments that are allowed to use databases and information systems in real time and upon
electronic request.
References:
1. Russell, S., & Norvig, P. (2020). Artificial Intelligence: A Modern Approach (4th ed.).
Pearson.
2. Mitchell, T. M. (1997). Machine Learning. McGraw-Hill..
3. Hastie, T., Tibshirani, R., & Friedman, J. (2009). The Elements of Statistical Learning:
Data Mining, Inference, and Prediction (2nd ed.). Springer.
4. Chandola, V., Banerjee, A., & Kumar, V. (2009). Anomaly detection: A survey. ACM
Computing Surveys (CSUR), 41(3), 1-58.
5. Bostrom, N., & Yudkowsky, E. (2014). The ethics of artificial intelligence. In
Cambridge Handbook of Artificial Intelligence. Cambridge University Press.
6. Yu, H., Berkovsky, S., & Castells, P. (Eds.). (2021). Machine Learning for Data Mining:
Basics and Applications. Elsevier.
