ARTIFICIAL INTELLIGENCE AND TAX FRAUD DETECTION: EMPIRICAL EVIDENCE FROM EDO STATE INTERNAL REVENUE SERVICE, NIGERIA

Authors

  • Dr. Tina Ashafoke Department of Accounting Faculty of Management Sciences University of Benin Benin City, Edo State, Nigeria. Author
  • Elugwu George Ugochukwu Department of Accounting Faculty of Management Sciences University of Benin Benin City, Edo State, Nigeria. Author
  • Obianke Osaretin Joy Department of Accounting Faculty of Management Sciences University of Benin Benin City, Edo State, Nigeria. Author

DOI:

https://doi.org/10.57233/ijamer.v2i3.010

Keywords:

Artificial Intelligence, Machine Learning, Predictive Analytics, Data Mining, Tax Fraud Detection, Tax Administration, Nigeria

Abstract

Tax fraud remains a major challenge to effective tax administration in Nigeria, as fraudulent reporting, concealment of taxable income and other forms of non-compliance can reduce government revenue and undermine the effectiveness of the tax system. Existing tax fraud detection mechanisms often rely on conventional audit procedures and manual examination of taxpayer records, which may be limited in their ability to process large volumes of data, identify complex fraud patterns and detect emerging forms of tax evasion. Against this background, this study examined the relationship between Artificial Intelligence (AI) and tax fraud detection in the Edo State Internal Revenue Service (EIRS), Nigeria. Specifically, the study examined the association between machine learning, predictive analytics, data mining and tax fraud detection. The study adopted a quantitative survey research design and obtained primary data from selected tax officers in EIRS. The data collected were analysed using multiple regression analysis. The findings revealed that machine learning had a positive and significant association with tax fraud detection (B = 0.276, t = 3.958, p = 0.000). Predictive analytics also had a positive and significant association with tax fraud detection (B = 0.259, t = 3.700, p = 0.000), while data mining had a positive and significant association with tax fraud detection (B = 0.312, t = 4.216, p = 0.000). The regression model produced an R of 0.842 and an R² of 0.709, indicating that 70.9% of the variation in tax fraud detection was jointly explained by machine learning, predictive analytics and data mining. The findings suggest that AI tools can enhance the efficiency of detecting abnormalities, strengthen audit effectiveness, and improve compliance monitoring and risk assessment. The study therefore recommends increased investment in AI-enabled tax administration technologies, continuous training of tax officers in the application of AI tools, and greater integration of machine learning, predictive analytics and data mining into tax fraud detection processes to improve tax compliance and revenue administration in Edo State.

Downloads

Download data is not yet available.

Author Biographies

  • Dr. Tina Ashafoke, Department of Accounting Faculty of Management Sciences University of Benin Benin City, Edo State, Nigeria.

    Department of Accounting
    Faculty of Management Sciences
    University of Benin
    Benin City, Edo State, Nigeria.

  • Elugwu George Ugochukwu, Department of Accounting Faculty of Management Sciences University of Benin Benin City, Edo State, Nigeria.

    Department of Accounting
    Faculty of Management Sciences
    University of Benin
    Benin City, Edo State, Nigeria.

  • Obianke Osaretin Joy, Department of Accounting Faculty of Management Sciences University of Benin Benin City, Edo State, Nigeria.

    Department of Accounting
    Faculty of Management Sciences
    University of Benin
    Benin City, Edo State, Nigeria.

Downloads

Published

2026-09-25

How to Cite

ARTIFICIAL INTELLIGENCE AND TAX FRAUD DETECTION: EMPIRICAL EVIDENCE FROM EDO STATE INTERNAL REVENUE SERVICE, NIGERIA. (2026). International Journal of Accounting, Management and Economic Review, 2(3), 194-216. https://doi.org/10.57233/ijamer.v2i3.010

Similar Articles

81-90 of 214

You may also start an advanced similarity search for this article.