Abstract
Tax authorities worldwide make extensive use of artificial intelligence (AI) technologies to automate various aspects of their tasks, such as answering taxpayer questions, assessing fraud risk, risk profiling, and auditing (selecting tax inspections). Since this automation has led to concerns about the impact of non-explainable AI systems on taxpayers’ rights, explainable AI (XAI) technologies appear to be fundamental for the lawful use of AI in the tax domain. This paper provides an initial map of the explainability requirements that AI systems must meet for tax applications. To this end, the paper examines the constitutional principles that guide taxation in democracies and the specific human rights system of the European Convention of Human Rights (ECHR), as interpreted by the European Court of Human Rights (ECtHR). Based on these requirements, the paper suggests how approaches to XAI might be deployed to address the specific needs of the various stakeholders in the tax domain.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In the broader context of public administratione, a recent report by the Netherlands Court of Audit analysed 9 AI systems used by the Dutch government, concluding that only 3 of them met minimum audit standards for governance and accountability, data and model management, and privacy protection: [14].
- 2.
A recent report from the Organisation for Economic Co-operation and Development (OECD) [8] used the term “Tax Administration 3.0” to mark a new stage of digitalization, in which taxation is moved closer to taxable events through built-in, automated compliance mechanisms and the interconnection between tax authority systems and the systems taxpayers use to run their businesses.
- 3.
Article 150, I, with a few exceptions presented in the same article As specified by the legislation implementing this constitutional provision, the prohibition encompasses any changes to the constitutive elements of a tax, such as the tax rate, the base, the triggering event or the applicable penalties (Article 97 National Tax Code).
- 4.
See, for example, the discussion of the Slovak constitutional case law below in this sub-section.
- 5.
This article deals with opacity stemming from technological factors, but opacity may also arise due to non-technological factors such as legislative opacity: [20].
- 6.
On opacity and its relation to algorithmic accountability, see [21].
- 7.
This judgment is in our view rightly considered by Philip Baker as one of the biggest failings of the ECtHR, since it practically means that under the ECHR, a taxpayer in member States of the ECHR does not have right “in an ordinary tax dispute to a fair trial by an independent and impartial tribunal” [32].
- 8.
Also, it is not unlikely that the ECtHR will revise the Ferrazzini case in the near future by allowing the application of Art. 6 of the ECHR to “normal” tax disputes, including those arising out of AI systems in tax law. In doing so, the Court would approach its treatment of tax disputes to the treatment of social security contributions, which are deemed to have private law features that outweigh the public elements of the obligation: [44]
- 9.
- 10.
The ECtHR indicates that the scope of protection under Art. 8 of the ECHR includes only personal data processing which concerns data regarding people’s private lives, or if data processing is extensive. Hence, not all personal data is covered by Art. 8 of the ECHR [54].
- 11.
A general prohibition of discrimination was enshrined in Article 1 of Protocol 12 of the ECHR. The protocol has already been ratified by enough signatories to come into effect, but nevertheless a considerable number of parties to the Convention have not ratified it.
- 12.
See, however, the discussion of [55] above.
- 13.
In most tax applications, AI systems deal with largely numerical data about relevant financial elements, which means that explanation approaches based on originally numerical features can play a crucial role. Nevertheless, some systems—such as those relying on text data or directly producing decisions that need to be grounded on legal arguments—might require the combination between explanation techniques and justification-based approaches for showing how the actions taken with basis on the explained outcomes can be sustained from the perspective of legal argumentation.
- 14.
Depending on what a particular stakeholder is tasked with doing, they are likely to require a different kind of knowledge to do it and, thus, to seek a different kind of explanation: [73].
- 15.
For a deeper discussion on this point, consult [75].
- 16.
By analogy to financial risk in loan risk assessment systems, as discussed in [73].
- 17.
Again, [75] can be consulted for more in-depth description and analysis.
References
Alm, J., Beebe, J., Kirsch, M.S., Marian, O., Soled, J.A.: New technologies and the evolution of tax compliance. Va. Tax Rev. 39, 287–356 (2020)
Di Puglia Pugliese, L., Guerriero, F., Macrina, G., Messina, E.: A natural language processing tool to support the electronic invoicing process in Italy. In: 2021 11th IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS), pp. 397–402 (2021). https://doi.org/10.1109/IDAACS53288.2021.9660987
Butler, J.: Analytical challenges in modern tax administration: a brief history of analytics at the IRS symposium on artificial intelligence & the future of tax law: AI in tax compliance and enforcement. Ohio St. Tech. L. J. 16, 258–277 (2020)
Lismont, J., et al.: Predicting tax avoidance by means of social network analytics. Decis. Support Syst. 108, 13–24 (2018). https://doi.org/10.1016/j.dss.2018.02.001
Antón, F.S.: Artificial intelligence and tax administration: strategy, applications and implications, with special reference to the tax inspection procedure. World Tax J. 13 (2021)
Hadwick, D., Lan, S.: Lessons to be learned from the dutch childcare allowance scandal: a comparative review of algorithmic governance by tax administrations in the Netherlands, France and Germany. World Tax J. 13 (2021)
Calo, R., Citron, D.K.: The automated administrative state: a crisis of legitimacy. Emory L. J. 70, 797–846 (2021)
OECD: Tax Administration 3.0: The Digital Transformation of Tax Administration. OECD, Paris (2020)
Braun Binder, N.: Artificial intelligence and taxation: risk management in fully automated taxation procedures. In: Wischmeyer, T., Rademacher, T. (eds.) Regulating Artificial Intelligence, pp. 295–306. Springer, Cham (2020). https://doi.org/10.1007/978-3-030-32361-5_13
Koops, B.-J.: The concept of function creep. Law Innov. Technol. 13, 29–56 (2021). https://doi.org/10.1080/17579961.2021.1898299
Scarcella, L.: Tax compliance and privacy rights in profiling and automated decision making. Internet Policy Review, vol. 8 (2019)
Sarra, C.: Put dialectics into the machine: protection against automatic-decision-making through a deeper understanding of contestability by design. Global Jurist, vol. 20 (2020). https://doi.org/10.1515/gj-2020-0003
Busuioc, M.: AI algorithmic oversight: new frontiers in regulation. In: Maggetti, M., Di Mascio, F., Natalini, A. (eds.) The Handbook on Regulatory Authorities. Edward Elgar Publishing, Cheltenham, Northampton (2022)
Rekenkamer, A.: An Audit of 9 Algorithms used by the Dutch Government. Netherlands Court of Audit, The Hague (2022)
Barredo, A., et al.: Explainable artificial intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI. Inf. Fusion. 58, 82–115 (2020). https://doi.org/10.1016/j.inffus.2019.12.012
Hattingh, J.: The multilateral instrument from a legal perspective: what may be the challenges? BFIT, vol. 71, (2017)
Huttner, L., Merigoux, D.: Catala: Moving Towards the Future of Legal Expert Systems (2022). https://doi.org/10.1007/s10506-022-09328-5
ECJ: Société d’investissement pour l’agriculture tropicale SA (SIAT) v État belge (Case C‑318/10) (2012)
ECJ: Itelcar — Automóveis de Aluguer Lda v Fazenda Pública (Case C‑282/12) (2013)
Burrell, J.: How the machine ‘thinks’: understanding opacity in machine learning algorithms. Big Data Soc. 3, 1–12 (2016). https://doi.org/10.1177/2053951715622512
Wieringa, M.: What to account for when accounting for algorithms: a systematic literature review on algorithmic accountability. In: Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, pp. 1–18. Association for Computing Machinery, Barcelona, Spain (2020). https://doi.org/10.1145/3351095.3372833
McConnell, S.: Code Complete. Microsoft Press, Redmond (2004)
Lawsky, S.B.: Formalizing the Code. Tax L. Rev. 70, 377–408 (2016)
Dijkstra, E.W.: The humble programmer. Commun. ACM. 15, 859–866 (1972). https://doi.org/10.1145/355604.361591
Horner, J.K., Symons, J.: Understanding error rates in software engineering: conceptual, empirical, and experimental approaches. Philos. Technol. 32(2), 363–378 (2019). https://doi.org/10.1007/s13347-019-00342-1
Richardson, R., Schultz, J.M., Southerland, V.M.: Litigating Algorithms 2019 US Report: New Challenges to Government Use of Algorithmic Decision Systems. AI Now Institute, New York (2019)
Slovenian Constitutional Court: Ústavného súdu Slovenskej republiky PL. ÚS 25/2019–117 V mene Slovenskej republiky (2021)
Harris, D.J., O’Boyle, M., Bates, E., Buckley, C.: Harris, O’Boyle & Warbrick: Law of the European Convention on Human Rights. Oxford University Press, Oxford (2014)
ECtHR: Practical Guide on Admissibility Criteria. European Court of Human Rights, Strasbourg (2022)
Emberland, M.: The Human Rights of Companies: Exploring the Structure of ECHR Protection. Oxford University Press, Oxford (2006)
ECtHR: Case of Ferrazzini v. Italy (Application no. 44759/98) (2001)
Baker, P.: 60 years of the European convention on human rights and taxation. European Taxation, vol. 61 (2021)
ECtHR: Case of Jussila v. Finland (Application no. 73053/01) (2006)
ECtHR: Case of Steininger v. Austria case (Application no. 21539/07) (2012)
ECtHR: Case of Chap Ltd v. Armenia case (Application no. 15485/09) (2017)
ECtHR: Case of Matyjek v. Poland (Application no. 38184/03) (2007)
ECtHR: Case of Moiseyev v. Russia (Application no. 62936/00) (2008)
ECtHR: Case of Mattoccia v. Italy (Application no. 23969/94) (2000)
ECJ: WebMindLicenses kft v Nemzeti Adó- és Vámhivatal Kiemelt Adó- és Vám Főigazgatóság (Case C-419/14) (2015)
ECtHR: Case of Ruiz Torija v. Spain (Application no. 18390/91) (1994)
ECtHR: Case of Fomin v. Moldova (Application no. 36755/06) (2011)
ECtHR: Case of Suominen v. Finland (Application no. 37801/97) (2003)
Dymitruk, M.: The right to a fair trial in automated civil proceedings. Masaryk Univ. J. Law Technol. 13, 27–44 (2019)
ECtHR: Case of Schouten and Meldrum v. the Netherlands (Application no. 19005/91; 19006/91) (1994)
Wieringa, M., van Schie, G., van de Vinne, M.: De discussie omtrent SyRI moet over meer dan alleen privacy gaan. https://ibestuur.nl/podium/de-discussie-omtrent-syri-moet-over-meer-dan-alleen-privacy-gaan. Accessed 03 Mar 2022
Henley, J., Booth, R.: Welfare surveillance system violates human rights, Dutch court rules (2020). https://www.theguardian.com/technology/2020/feb/05/welfare-surveillance-system-violates-human-rights-dutch-court-rules
Simonite, T.: Europe limits government by algorithm. The US, Not So Much (2021). https://www.wired.com/story/europe-limits-government-algorithm-us-not-much/
Rechtbank den Haag: NJCM et al v, Netherlands (2020)
van Bekkum, M., Borgesius, F.Z.: Digital welfare fraud detection and the Dutch SyRI judgment. Eur. J. Soc. Secur. 23(4), 323–340 (2021). https://doi.org/10.1177/13882627211031257
ECtHR: Case of S. and Marper v. The United Kingdom (Applications nos. 30562/04 and 30566/04) (2008)
Gellert, R., van Bekkum, M., Zuiderveen Borgesius, F.: The Ola & Uber judgments: for the first time a court recognises a GDPR right to an explanation for algorithmic decision-making. https://eulawanalysis.blogspot.com/2021/04/the-ola-uber-judgments-for-first-time.html. Accessed 30 Apr 2021
Rechtbank Amsterdam: “Uber employment” (HA20 — RK258) (2021)
Rechtbank Amsterdam: “Ola” (HA20 — RK207) (2021)
De Hert, P., Gutwirth, S.: Data protection in the case law of strasbourg and luxemburg: constitutionalisation in action. In: Gutwirth, S., Poullet, Y., De Hert, P., de Terwangne, C., Nouwt, S. (eds.) Reinventing Data Protection? pp. 3–44. Springer Netherlands, Dordrecht (2009). https://doi.org/10.1007/978-1-4020-9498-9_1
ECtHR: Case of I v. Finland (Application no. 20511/03) (2008)
Lagioia, F., Sartor, G.: Artificial intelligence in the big data era: risks and opportunities. In: Cannatacci, J., Falce, V., Pollicino, O. (eds.) Legal Challenges of Big Data, pp. 280–307. Edward Elgar, Northampton (2020)
Lessig, L.: Law Regulating Code Regulating Law. Loy. U. Chi. L.J. 35, 1–14 (2003)
Artosi, A.: Technical normativity. In: Chiodo, S., Schiaffonati, V. (eds.) Italian Philosophy of Technology: Socio-Cultural, Legal, Scientific and Aesthetic Perspectives on Technology, pp. 149–160. Springer International Publishing, Cham (2021). https://doi.org/10.1007/978-3-030-54522-2_10
Diver, L.: Digisprudence: Code as Law Rebooted. Edinburgh University Press, Edinburgh (2021)
ECtHR: Guide on Article 14 of the European Convention on Human Rights and on Article 1 of Protocol No. 12 to the Convention: Prohibition of discrimination. European Court of Human Rights, Strasbourg (2021)
ECtHR: Case of Darby v. Sweden (Application no. 11581/85) (1990)
ECtHR: Case of Glor v. Switzerland (Application no. 13444/04) (2009)
ECtHR: Case of Guberina v. Croatia (Application no. 23682/13) (2016)
ECtHR: Taxation and the european convention on human rights. European Court of Human Rights, Strasbourg (2021)
Goodman, B.W.: Economic models of (algorithmic) discrimination. In: Presented at the 29th Conference on Neural Information Processing Systems (NIPS 2016), Barcelona (2016)
Muller, C.: The Impact of Artificial Intelligence on Human Rights, Democracy and the Rule of Law. Council of Europe, Strasbourg (2020)
Kuźniacki, B., Tyliński, K.: Identifying the potential and risks of integration of AI to taxation: the case of general anti-avoidance rule. In: D’Agostino, G., Gaon, A., Piovesan, C. (eds.) Leading Legal Disruption: Artificial Intelligence and a Toolkit for Lawyers and the Law. Carswell, Toronto (2021)
Humphreys, P.: The philosophical novelty of computer simulation methods. Synthese 169, 615–626 (2009). https://doi.org/10.1007/s11229-008-9435-2
Durán, J.M., Formanek, N.: Grounds for trust: essential epistemic opacity and computational reliabilism. Mind. Mach. 28(4), 645–666 (2018). https://doi.org/10.1007/s11023-018-9481-6
Steging, C., Renooij, S., Verheij, B.: Rationale discovery and explainable AI. In: Legal Knowledge and Information Systems, pp. 225–234. IOS Press, Vilnius (2021). https://doi.org/10.3233/FAIA210341
Mumford, J., Atkinson, K., Bench-Capon, T.: Machine learning and legal argument. In: Schweighofer, E. (ed.) Legal Knowledge and Information Systems, pp. 191–196. IOS Press, Vilnius (2021)
Górski, Ł., Ramakrishna, S.: Explainable artificial intelligence, lawyer’s perspective. In: Proceedings of the Eighteenth International Conference on Artificial Intelligence and Law, pp. 60–68. Association for Computing Machinery, New York, NY, USA (2021). https://doi.org/10.1145/3462757.3466145
Zednik, C.: Solving the black box problem: a normative framework for explainable artificial intelligence. Philos. Technol. 34(2), 265–288 (2019). https://doi.org/10.1007/s13347-019-00382-7
Mehdiyev, N., Houy, C., Gutermuth, O., Mayer, L., Fettke, P.: Explainable artificial intelligence (XAI) supporting public administration processes – on the potential of XAI in tax audit processes. In: Ahlemann, F., Schütte, R., Stieglitz, S. (eds.) WI 2021. LNISO, vol. 46, pp. 413–428. Springer, Cham (2021). https://doi.org/10.1007/978-3-030-86790-4_28
Molnar, C.: Interpretable Machine Learning. Lulu.com (2020)
Acknowledgements
The authors would like to thank Anca Radu, Réka Markovich, and three anonymous EXTRAAMAS reviewers for their feedback on this paper. The authors, however, bear full responsibility for the paper.
All authors acknowledge that this paper is based upon work supported in whole by The Notre Dame-IBM Tech Ethics Lab. Such support does not constitute an endorsement by the sponsor of the authors’ views expressed in this publication.
Błażej Kuźniacki acknowledges that his work on this paper has been developed within the framework of the Amsterdam Centre for Tax Law (ACTL) research project “Designing the tax system for a cashless, platform-based and technology-driven society” (CPT project). The CPT project is financed with University funding and funds provided by external stakeholders (i.e. businesses and governments) interested in supporting academic research to design fair, efficient and fraud-proof tax systems. For more information about the CPT project and its partners, please visit its website https://actl.uva.nl/cpt-project/cpt-project.html. The support received by the author within the framework of the CPT project does not constitute an endorsement by the sponsors of the views expressed in this publication by the author.
Marco Almada would also like to thank Fundacion Carolina for granting him a doctoral scholarship, under which his work on this project was partially funded.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 The Author(s), under exclusive license to Springer Nature Switzerland AG
About this paper
Cite this paper
Kuzniacki, B., Almada, M., Tyliński, K., Górski, Ł. (2022). Requirements for Tax XAI Under Constitutional Principles and Human Rights. In: Calvaresi, D., Najjar, A., Winikoff, M., Främling, K. (eds) Explainable and Transparent AI and Multi-Agent Systems. EXTRAAMAS 2022. Lecture Notes in Computer Science(), vol 13283. Springer, Cham. https://doi.org/10.1007/978-3-031-15565-9_14
Download citation
DOI: https://doi.org/10.1007/978-3-031-15565-9_14
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-031-15564-2
Online ISBN: 978-3-031-15565-9
eBook Packages: Computer ScienceComputer Science (R0)