AI for Tax and Customs Anomaly Detection: A PRISMA-Guided Systematic Review and Proposed Hybrid AI Framework for Trade and Revenue Administration in Emerging Digital Economies

References

[1] [Alexopoulos-2025] Alexopoulos, A., Dellaportas, P., Gyoshev, S. B., Kotsogiannis, C., Olhede, S. C., & Pavkov, T. (2025). A network approach to detect value added tax fraud. Journal of the Royal Statistical Society Series A: Statistics in Society, 188(3). https://doi.org/10.1093/jrsssa/qnaf205

[2] [Alsadhan-2023] Alsadhan, N. (2023). A multi-module machine learning approach to detect tax fraud. Computer Systems Science and Engineering, 46(1), 241–253. https://doi.org/10.32604/csse.2023.033375

[3] [Alwanin-2025] Alwanin, R., Ben Ismail, M. M., & Bchir, O. (2025/2026). Customs fraud detection using a gradient boosting approach for joint classification and risk estimation. Scientific Reports, 16(1), 3432. https://doi.org/10.1038/s41598-025-33382-z

[4] [APEC-2025] APEC. (2025). Application of data analytics in customs risk management. Asia-Pacific Economic Cooperation. Available: https://www.apec.org.

[5] [Arora-2024] Arora, S., Rani, R., & Saxena, N. (2024). A systematic review on detection and adaptation of concept drift in streaming data using machine learning techniques. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 14, e1536. https://doi.org/10.1002/widm.1536

[6] [Belahouaoui-2025] Belahouaoui, R., & Alm, J. (2025). Tax fraud detection using artificial intelligence-based technologies: Trends and implications. Journal of Risk and Financial Management, 18(9), 502. https://doi.org/10.3390/jrfm18090502

[7] [CIAT-2021] CIAT. (2021). Explainable artificial intelligence (XAI) and its importance in tax administration. Inter-American Center of Tax Administrations. Available: https://www.ciat.org.

[8] [Dangsawang-2024] Dangsawang, B., & Nuchitprasitchai, S. (2024). A machine learning approach for detecting customs fraud through unstructured data analysis in social media. Decision Analytics Journal, 10, 100408. https://doi.org/10.1016/j.dajour.2024.100408

[9] [Hovakimyan-2024] Hovakimyan, G., & Bravo, J. (2024). Evolving strategies in machine learning: A systematic review of concept drift detection methods. Information, 15(12), 786. https://doi.org/10.3390/info15120786

[10] [Karklina-2024] Karklina-Admine, S., Cevers, A., Kovalenko, A., & Auzins, A. (2024). Challenges for customs risk management today: A literature review. Journal of Risk and Financial Management, 17(8), 321. https://doi.org/10.3390/jrfm17080321

[11] [Kim-2020] Kim, S., Tsai, Y.-C., Singh, K., Cha, M., Li, C.-T., Lin, S.-D., & Kim, Y.-R. (2020). DATE: Dual attentive tree-aware embedding for customs fraud detection. Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 2880–2890). ACM. https://doi.org/10.1145/3394486.3403339

[12] [Kim-2023] Kim, S., Mai, T.-D., Han, S., Park, S., Nguyen Duc Khanh, T., So, J., Singh, K., & Cha, M. (2023). Active learning for human-in-the-loop customs inspection. IEEE Transactions on Knowledge and Data Engineering, 35(12), 12039–12050. https://doi.org/10.1109/TKDE.2022.3144299

[13] [Kuzniacki-2022] Kuźniacki, B., Almada, M., Tylikński, K., Górski, Ł., Winogradska, B., & Zeldenrust, R. (2022). Towards explainable artificial intelligence (XAI) in tax law: The need for a minimum legal standard. World Tax Journal, 14(4). https://doi.org/10.59403/2yhh9pa

[14] [Li-2023] Li, Z., Zhu, Y., & van Leeuwen, M. (2023). A survey on explainable anomaly detection. ACM Transactions on Knowledge Discovery from Data, 18(1). https://doi.org/10.1145/3609333

[15] [Malashin-2025] Malashin, I. P., Masich, A., Kolodyazhny, Y., & Roas, S. (2025). Minimizing unnecessary tax audits using multi-objective optimization and machine learning hyperparameter tuning. Frontiers in Artificial Intelligence, 8, 1669191. https://doi.org/10.3389/frai.2025.1669191

[16] [Mittra-2026] Mittra, S. (2026). Harnessing AI and data for tax administration. International Growth Centre. Available: https://www.theigc.org.

[17] [Murorunkwere-2022] Murorunkwere, B. F., Tuyishimire, O., Haughton, D., & Nzabanita, J. (2022). Fraud detection using neural networks: A case study of income tax. Future Internet, 14(6), 168. https://doi.org/10.3390/fi14060168

[18] [Murorunkwere-2023] Murorunkwere, B. F., Haughton, D., Nzabanita, J., & Tuyishimire, O. (2023). Predicting tax fraud using supervised machine learning approach. African Journal of Science, Technology, Innovation and Development, 15(6), 731–742. https://doi.org/10.1080/20421338.2023.2187930

[19] [Nassif-2021] Nassif, A. B., Talib, M. A., Nasir, Q., & Dakalbab, F. M. (2021). Machine learning for anomaly detection: A systematic review. IEEE Access, 9, 78658–78700. https://doi.org/10.1109/ACCESS.2021.3083060

[20] [Nelson-2020] Nelson, C. (2020). Machine learning for detection of trade in strategic goods: An approach to support future customs enforcement and outreach. World Customs Journal, 14(2), 119–130. https://doi.org/10.55596/001c.116422

[21] [OECD-2020] OECD. (2020). Tax administration 3.0: The digital transformation of tax administration. OECD Publishing. https://doi.org/10.1787/ca274cc5-en

[22] [OECD-2022a] OECD. (2022a). Tax administration 3.0 and the digital identification of taxpayers: Initial findings. OECD Publishing. https://doi.org/10.1787/3ab1789a-en

[23] [OECD-2022b] OECD. (2022b). Tax administration 3.0 and electronic invoicing: Initial findings. OECD Publishing. https://doi.org/10.1787/2ffc88ed-en

[24] [OECD-2025a] OECD. (2025a). Tax administration 2025: Comparative information on OECD and other advanced and emerging economies. OECD Publishing. Available: https://www.oecd.org/en/publications/tax-administration-2025_88538e8b-en.html

[25] [OECD-2025b] OECD. (2025b). Tax administration digitalisation and digital transformation initiatives. OECD Publishing. https://doi.org/10.1787/c076d776-en

[26] [OECD-2025c] OECD. (2025c). Tax administration 3.0: From vision to strategy. OECD Publishing. Available: https://www.oecd.org/tax/forum-on-tax-administration.

[27] [OECD-2025d] OECD. (2025d). AI in tax administration. In Governing with artificial intelligence: The state of play and way forward in core government functions. OECD Publishing. https://doi.org/10.1787/795de142-en

[28] [OECD-2025e] OECD. (2025e). The digitalisation of trade documents and processes: Going paperless today, going paperless tomorrow. OECD Trade Policy Papers, 297. OECD Publishing. https://doi.org/10.1787/64872f25-en

[29] [OECD-2025f] OECD. (2025f). Governing with artificial intelligence: The state of play and way forward in core government functions. OECD Publishing. https://doi.org/10.1787/795de142-en

[30] [OECD-2024] OECD, CIAT, IOTA, IMF, & partners. (2024). Inventory of tax technology initiatives. OECD Forum on Tax Administration. Available: https://www.oecd.org/tax/forum-on-tax-administration/tax-technology-tools-and-digital-solutions.

[31] [Page-2021a] Page, M. J., McKenzie, J. E., Bossuyt, P. M., et al. (2021a). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Journal of Clinical Epidemiology, 134, 178–189. https://doi.org/10.1016/j.jclinepi.2021.03.001

[32] [Page-2021b] Page, M. J., Moher, D., Bossuyt, P. M., et al. (2021b). PRISMA 2020 explanation and elaboration: Updated guidance and exemplars for reporting systematic reviews. BMJ, 372, n160. https://doi.org/10.1136/bmj.n160

[33] [Ruvalcaba-2025] Ruvalcaba-Gómez, E. A., & García-Benitez, V. H. (2025). Governance of artificial intelligence in the public sector: Analysis of public policies in Spain and Mexico. Review of Policy Research, 42(4). https://doi.org/10.1111/ropr.70057

[34] [Singh-2023] Singh, K., Tsai, Y.-C., Li, C.-T., Cha, M., & Lin, S.-D. (2023). GraphFC: Customs fraud detection with label scarcity. Proceedings of the 32nd ACM International Conference on Information and Knowledge Management. ACM. https://doi.org/10.1145/3583780.3614690

[35] [UNDP-2025] UNDP. (2025). AI for the next generation of public services: Global trends and case studies. United Nations Development Programme. Available: https://www.undp.org/publications/ai-next-generation-public-services.

[36] [Vanhoeyveld-2020] Vanhoeyveld, J., Martens, D., & Peeters, B. (2020). Value-added tax fraud detection with scalable anomaly detection techniques. Applied Soft Computing, 86, 105895. https://doi.org/10.1016/j.asoc.2019.105895

[37] [Vijayakumar-2025] Vijayakumar, S. (2025). Technology-centric and data-driven customs risk management for supply chain security. World Customs Journal, 19(1), 38–63. https://doi.org/10.55596/001c.131745

[38] [WCO-2023] WCO. (2023). Compendium of case studies on e-commerce (4th ed.). World Customs Organization. Available: https://www.wcoomd.org.

[39] [ZATCA-2023] Zakat, Tax and Customs Authority (ZATCA). (2023). E-invoicing (Fatoorah) programme: Phase 2 integration requirements and technical specifications. ZATCA, Riyadh, Kingdom of Saudi Arabia. Available: https://zatca.gov.sa/en/E-Invoicing/Pages/default.aspx.

[40] [WCO-WTO-2022] WCO & WTO. (2022). Study report on disruptive technologies. World Customs Organization and World Trade Organization. Available: https://www.wcoomd.org.

[41] [Zheng-2024] Zheng, Q., Xu, Y., Liu, H., Shi, B., Wang, J., & Dong, B. (2024). A survey of tax risk detection using data mining techniques. Engineering, 34, 43–57. https://doi.org/10.1016/j.eng.2023.07.014

@article{1715244,
      author = {Obaid ur Rehman Qureshi},
      title = {AI for Tax and Customs Anomaly Detection: A PRISMA-Guided Systematic Review and Proposed Hybrid AI Framework for Trade and Revenue Administration in Emerging Digital Economies},
      journal = {Iconic Research And Engineering Journals},
      year = {2026},
      volume = {9},
      number = {9},
      pages = {1526-1541},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1715244.pdf},
      abstract = {Tax and customs administrations in emerging digital economies face simultaneous pressure to accelerate trade clearance, protect revenue, and maintain procedural fairness under rapidly digitising commercial activity. Traditional rule-based targeting, while operationally important, is increasingly outpaced by cross-border e-commerce growth, valuation manipulation, synthetic invoicing, shell entity proliferation, and adaptive fraud behaviour. This review synthesises peer-reviewed literature and high-value institutional publications from 2020 to early 2026 on artificial intelligence (AI), machine learning (ML), graph analytics, anomaly detection, explainable AI (XAI), and human-in-the-loop (HITL) risk management applied to tax and customs declarations. Following PRISMA 2020 reporting principles, 38 principal sources were selected from 518 initially identified records across Scopus, Web of Science, IEEE Xplore, ACM Digital Library, SSRN, Google Scholar, and institutional repositories of the OECD, WCO, APEC, CIAT, UNDP, and the International Growth Centre. The synthesis reveals three structural findings. First, high-performing anomaly detection systems consistently employ hybrid architectures that integrate supervised fraud scoring, unsupervised or semi-supervised anomaly detection, graph-based relational modelling, and active learning — rather than any single model family. Second, performance claims are highly sensitive to label scarcity, inspection bias, concept drift, and operational constraints including audit budgets, explainability mandates, and workflow integration requirements. Third, governance, legal explainability, human oversight, and continuous model monitoring are necessary design requirements — not post-implementation considerations — for public-sector deployment. Based on this evidence, the paper proposes a seven-layer AI-driven anomaly detection framework specifically designed for revenue administrations in emerging digital economies. The framework integrates data identity and interoperability foundations, risk-feature engineering, hybrid model stacking, explainable risk presentation, human-in-the-loop case management, continuous learning and drift control, and institutional governance and recourse mechanisms. The framework is directly applicable to ZATCA (Zakat, Tax and Customs Authority) in Saudi Arabia, GCC customs administrations, and comparable digital revenue administrations across emerging economies. The review concludes that technically credible AI-enabled anomaly detection is now achievable, but that scaled and sustainable deployment depends less on model novelty than on data quality, interoperable digital infrastructure, inspection feedback loops, legal explainability standards, and governance maturity.},
      keywords = {Customs Fraud Detection; Tax Anomaly Detection; AI In Government; PRISMA Systematic Review; Emerging Digital Economies; Explainable AI; Graph Neural Networks; VAT Fraud; ZATCA; Vision 2030; Hybrid Detection Architecture; Human-In-The-Loop; Concept Drift; Active Learning.},
      month = {March},
      doi = {https://doi.org/10.64388/IREV9I9-1715244}
  }