REVISING THE TAX AUDIT: ANALYSING THE HUMAN FACTOR AS A THIRD PARTY CONTROLLER IN ALGORITHMIC DECISION-MAKING PROCESSES

Yıl 2024, Sayı: 31, 47 - 58, 01.12.2024

Ayşegül Yücel , Burçin Bozdoğanoğlu

https://doi.org/10.58348/denetisim.1540801

Öz

Artificial intelligence technology is a digital tool that can learn, analyse and make decisions based on processed data and codes, which has started to be used in the workflows of the public and private sectors in order to save time and cost. Artificial intelligence technology, which is capable of deep and continuous learning, can make decisions on important issues such as justice, defence, health and law depending on its algorithms, and can make predictions based on the data it processes and classifies. However, it is possible that unlawful consequences may arise due to the non-transparency of the data sets and codes on which artificial intelligence technology makes decisions and on what grounds. In such a case, artificial intelligence users are likely to be harmed by the legal and actual effects of decisions taken as a result of non-transparent processes. This situation brings up the necessity of supervising artificial intelligence, which is a human product, at the design and implementation stages in order to prevent possible violations of rights. However, the inadequacy of the legal legislation on the supervision of artificial intelligence technology and the uncertainty of which type of supervision will be applied at which stage causes a great deficiency in the literature. In this study, the methods and ways of auditing artificial intelligence technology are discussed, and it is stated that the traditional audit methods have lost their effect in the digital world. It is thought that third-party auditing, which is a new type of auditing, will be more effective in AI auditing, especially if it uses the ethics-based method. After discussing artificial intelligence auditing from a general perspective, the study calls for a revision of the tax audit process in order to prevent violations of rights caused by artificial intelligence technology used in the taxation process, and proposes a hybrid audit model in which traditional audit models and third-party audits work in cooperation for an uninterrupted and continuous tax management.

Anahtar Kelimeler

Tax Audit, Artificial Intelligence, Algorithmic Decision Making, Black Box, Third-Party Audit

VERGİ DENETİMİNİ REVİZE ETMEK: ALGORİTMİK KARAR ALMA SÜREÇLERİNDE ÜÇÜNCÜ TARAF KONTROLÖRÜ OLARAK İNSAN FAKTÖRÜNÜN İNCELENMESİ

Öz

Yapay zeka teknolojisi, kamu ve özel sektörün iş akışlarında zaman ve maliyetten tasarruf etme amacıyla kullanılmaya başlayan, işlenen veri ve kodlara bağlı öğrenebilen, analiz edebilen, karar alabilen dijital bir araçtır. Derin ve sürekli öğrenme yetisine sahip yapay zeka teknolojisi kullanıldığı adalet, savunma, sağlık, hukuk gibi önemli konularda algoritmalarına bağlı olarak kararlar alabilmekte, işlediği ve tasnif ettiği veriler sayesinde tahmine dayalı işlemler yapabilmektedir. Fakat yapay zeka teknolojisinin kararları hangi veri setleri ve kodlara dayanarak ve hangi gerekçelerle aldığının şeffaf olmayışı nedeniyle hukuka aykırı sonuçların doğması muhtemeldir. Böyle bir durumda yapay zeka kullanıcıları şeffaf olmayan süreçler sonucunda alınan kararların hukuki ve fiili etkilerinden zarar görme ihtimali taşımaktadır. Bu durum, yaşanabilecek hak ihlallerinin önüne geçmek amacıyla, insan ürünü olan yapay zekanın tasarım ve uygulama aşamalarında denetlenmesi gerekliliğini gündeme getirmektedir. Fakat yapay zeka teknolojisinin denetlenmesine dair hukuki mevzuatın yetersizliği, hangi denetim türünün hangi aşamada uygulacağının dahi belirsizliği literatürde büyük bir eksikliğe neden olmaktadır. Çalışmada, yapay zeka teknolojisinin hangi yöntem ve yollar izlenerek denetleneceği tartışılmakta, geleneksel denetim yollarının dijital dünyada etkisinin kaybolduğu belirtilmektedir. Yeni bir denetim türü olan üçüncü taraf denetiminin özellikle etik temelli yöntemi kullanmasının yapay zeka denetiminde daha etkili olacağı düşünülmektedir. Çalışma, yapay zeka denetimini genel bir bakış açısıyla ele aldıktan sonra, vergilendirme sürecinde kullanılan yapay zeka teknolojisinin neden olacağı hak ihlallerinin önüne geçilmesi amacıyla vergi denetim sürecine revizyon talebinde bulunmakta, kesintisiz ve sürekli bir vergi yönetimi açısından geleneksel denetim modelleri ile üçüncü taraf denetimlerinin işbirliği içinde çalıştığı karma bir denetim modeli önermektedir.

Anahtar Kelimeler

Vergi Denetimi, Yapay Zekâ, Algoritmik Karar Alma, Kara Kutu, Üçüncü Taraf Denetimi

Kaynakça

• Adelekan O. A., Adisa O., Ilugbusi B. S., Obi O. C., Awonuga K. F., Asuzu O. F. & Ndubuisi N. L. (2024). Evolving Tax Compliance In The Digital Era: A Comparative Analysis Of Ai-Driven Models And Blockchain Technology In U.S. Tax Administration. Computer Science. IT Research Journal, 5(2), 311-335.
• Alon-Barkat S., Busuioc M. (2023). Human–AI Interactions in Public Sector Decision Making: “Automation Bias” and “Selective Adherence” to Algorithmic Advice. Journal of Public Administration Research and Theory. 33(1), 153-169.
• Auld G., Casovan A., Clarke A. & Faveri B. (2022) Governing AI through Ethical Standards: Learning From the Experiences of Other Private Governance Initiatives. Journal of European Public Policy, 29(11), 1822-1844.
• Bal, A. (2019). Ruled by Algorithms: The Use of ‘Black Box’ Models in Tax Law. Tax Notes International, 95(12), 1158-1165.
• Bozdoğanoğlu, B. (2024). Vergi İdarelerinde Yapay Zekâ Sistemlerinin Kullanımının Riskleri: AB Uygulamaları ve Mevzuatı Çerçevesinde Değerlendirmeler, Malî Hukuk Dergisi, 20(234), 2024, 1003 -1062.
• Bozkurt, B. (2016). Denetim Kavrami Ve Denetim Anlayişindaki Gelişmeler. Denetişim, (12), 56-62.
• Chaudhary, G. (2024). Unveiling the Black Box: Bringing Algorithmic Transparency to AI. Masaryk University Journal of Law and Technology, 18(1), 93-122.
• Faveri B. & Auld G. (2023). Informing Possible Futures for the use of Third-Party Audits in AI Regulations. Carleton University, School of Public Policy and Administration. https://repository.library.carleton.ca/concern/research_works/2z10wr54f?locale=en. (Erişim Tarihi, 21.08.2024).
• Hartmann D., Renato Laranjeira de Pereira J., Streitbörger C. & Berendt B. (2024). Addressing the Regulatory Gap: Moving Towards an EU AI Audit Ecosystem Beyond the AIA by Including Civil Society. https://arxiv.org/html/2403.07904v1. (Erişim Tarihi, 25.08.2024).
• Huang, Z. (2018). Discussion on the Development of Artificial Intelligence in Taxation. American Journal of Industrial and Business Management, 8, 1817-1824.
• Kuźniacki, B., Almada, M., Tyliński, 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), 1-28.
• Mökander, J. (2023). Auditing of AI: Legal, Ethical and Technical Approaches. Digital Society. 2(49), https://link.springer.com/article/10.1007/s44206-023-00074-y.
• Mökander J., Axente M. (2023). Ethicsbased Auditing of Automated DecisionMaking Systems: Intervention Points and Policy Implications. AI Society (38), 153-171.
• Mökander J., Curl J. & Kshirsagar M. (2024). A Blueprint for Auditing Generative AI. https://www.researchgate.net/publication/382080223_A_Blueprint_for_Auditing_Generative_AI. (Erişim Tarihi, 05.08.2024).
• OECD. (2023). Tax Administration 2023: Comparative Information on OECD and other Advanced and Emerging Economies. Paris: OECD Publishing.
• Pica, L.M. (2022) Artificial Intelligence, Tax Law and (Intelligent?) Tax Administration. European Review of Digital Administration & Law, 3(1), 141-149.
• Raji I. N., Xu P., Honigsberg C., Ho D. (2022). Outsider Oversight: Designing a Third Party Audit Ecosystem for AI Governance. In Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society (AIES’22), 1–3, Oxford, United Kingdom.
• Rinta-Kahila T., Someh I., Gillespie N., Indulska M. & Gregor S. (2022) Algorithmic Decision-Making And System Destructiveness: A Case Of Automatic Debt Recovery, European Journal of Information Systems, 31(3), 313-338.
• William, R. (2022). Rethinking Administrative Law for Algorithmic Decision Making. Oxford Journal of Legal Studies, 42(2), 468–494.
• Zaqeebaa N., Alqudaha H., Alshira’h A. F., Lutfi A., Almaiah M. A. & Alrawad M. (2024). The Impact of Using Types of Artificial Intelligence Technology in Monitoring Tax Payments. International Journal of Data and Network Science, 8, 1577–1586.
• Anthropic (2024). Third-Party Testing As A Key Ingredient Of AI Policy. https://www.anthropic.com/news/third-party-testing. (Erişim Tarihi, 15.08.2024).
• Information Commissioner's Office (ICO) (2022, Ekim). What Is Automated Individual Decision-Making And Profiling?. https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/individual-rights/automated-decision-making-and-profiling/what-is-automated-individual-decision-making-and-profiling/. (Erişim Tarihi, 12.08.2024).
• Information Commissioner's Office (ICO) (2023, Kasım). Ethics, Transparency and Accountability Framework for Automated Decision-Making?. https://www.gov.uk/government/publications/ethics-transparency-and-accountability-framework-for-automated-decision-making/ethics-transparency-and-accountability-framework-for-automated-decision-making. (Erişim Tarihi, 10.08.2024).