BEYOND DETERMINISM – LEGAL RESPONSIBILITY IN THE AGE OF PROBABILISTIC AI SYSTEMS

Žaklina Spalević; Ratko Ivković; Milan Ječmenić

doi:10.5937/ptp2603134S

Authors

Žaklina Spalević Singidunum University, Faculty of Tourism and Hospitality Management, Belgrade, Serbia https://orcid.org/0000-0001-8502-2038
Ratko Ivković State University of Novi Pazar, Department of Technical Sciences, Novi Pazar, Serbia https://orcid.org/0000-0002-6557-4553
Milan Ječmenić MB University Belgrade, Business and Law Faculty, Belgrade, Serbia https://orcid.org/0009-0003-6109-3490

DOI:

https://doi.org/10.5937/ptp2603134S

Keywords:

artificial intelligence, non-determinism, law of obligations, EU AI Act, consumer protection

Abstract

This paper examines the fundamental distinction between classical deterministic programming and contemporary artificial intelligence systems, with particular emphasis on the legal consequences arising from the non-deterministic behavior of AI models. By analyzing the basic postulates of programming, the Softmax function as a key mechanism for output generation, and stochastic factors in training and generative processes, the paper demonstrates that artificial intelligence deviates from expected norms of predictability—an aspect that has far-reaching implications for the law of obligations, consumer protection, and tort liability regimes. In addition to a theoretical overview, the paper provides a legal analysis of the European Union’s AI Act in the context of non-determinism, highlighting regulatory efforts to address this characteristic of AI models through a risk-classification system as well as technical and organizational requirements. The paper aims to contribute to an interdisciplinary understanding of AI-related challenges by positioning legal scholarship as a key factor in developing a new model of liability in the digital age.

References

Atil, B., Aykent, S., Chittams, A., Fu, L., & Passonneau, R. J. (2025). Non-Determinism of “Deterministic” LLM Settings. arXiv preprint arXiv:2408.04667v5. DOI: 10.48550/arXiv.2408.04667

Devetzis, D., & Samaras, S. (2024). Consumer Protection Safeguards after the AI Act. Perspectives of Law and Public Administration, 13(2), pp. 298–309. DOI: 10.62768/PLPA/2024/13/2/12

Directive (EU) 2019/770 of the European Parliament and of the Council of 20 May 2019 on certain aspects concerning contracts for the supply of digital content and digital services. Official Journal of the European Union (OJ). Downloaded 2026, January 20, from https://eur-lex.europa.eu/eli/dir/2019/770/oj/eng

Directive (EU) 2024/2853 of the European Parliament and of the Council of 23 October 2024 on liability for defective products and repealing Council Directive 85/374/EEC (Product Liability Directive). Official Journal of the European Union (OJ). Downloaded 2026, January 26, from https://eur-lex.europa.eu/eli/dir/2024/2853/oj/eng

Drabent, W. (2025). On systematic construction of correct logic programs. arXiv preprint arXiv:2508.16782. DOI: 10.48550/arXiv.2508.16782

EU Artificial Intelligence Act. (2024). The EU Artificial Intelligence Act: Up-to-date developments and analyses of the EU AI Act. Downloaded 2025, December 30, from https://artificialintelligenceact.eu/

European Commission. (2024). Proposal for a Directive on Adapting NonContractual Civil Liability Rules to Artificial Intelligence (COM (2024)). Downloaded 2026, January 20, from https://www.europarl.europa.eu/RegData/etudes/STUD/2024/762861/EPRS_STU(2024)762861_EN.pdf

European Parliament. (2025). AI liability directive – Legislative Train Schedule (includes reference to withdrawal notice in OJ C/2025/5423, 6 October 2025). Downloaded 2026, January 26, from https://www.europarl.europa.eu/legislative-train/theme-a-europe-fit-for-the-digital-age/file-ai-liability-directive

European Parliamentary Research Service. (2025). Artificial Intelligence and Civil Liability – A European Perspective. Downloaded 2026, January 21, from https://www.europarl. europa.eu/RegData/etudes/STUD/2025/776426/IUST_STU(2025)776426_EN.pdf

Franke, M., & Degen, J. (2023). The softmax function: Properties, motivation, and interpretation. PsyArXiv. DOI: 10.31234/osf.io/vsw4711. Fu, T. (2026). Beyond Reproducibility: Token Probabilities Expose Large Language Model Nondeterminism. arXiv preprint arXiv:2601.06118v1. DOI: 10.48550/arXiv.2601.06118

Gasparri, W. (2025). Artificial intelligence and AI Act: from the individual to the algorithm? Zbornik radova Pravnog fakulteta u Novom Sadu, 59(1), pp. 291–314. DOI: 10.5937/zrpfns59-57519

Hagen, E. H., Chater, N., Gallistel, C. R., Houston, A., Kacelnik, A., Kalenscher, T., Nettle, D., Oppenheimer, D., & Stephens, D. W. (2012). Decision making: What can evolution do for us? In: Hammerstein P., & Stevens, J., R. (eds.), Evolution and the mechanisms of decision making (pp. 97–126). MIT Press

Herrera-Poyatos, D., Peláez-González, C., Zuheros, C., Herrera-Poyatos, A., Tejedor, V., Herrera, F., & Montes, R. (2025). An overview of model uncertainty and variability in LLM-based sentiment analysis: challenges, mitigation strategies, and the role of explainability. Frontiers in Artificial Intelligence, 8, 1609097. DOI: 10.3389/frai.2025.1609097

ISO/IEC. (2023). ISO/IEC 23894:2023 – Artificial intelligence – Guidance on risk management. International Organization for Standardization (ISO). Downloaded 2026, January 27, from https://www.iso.org/standard/77304.html

ISO/IEC. (2023). ISO/IEC 42001:2023 – Artificial intelligence – Management system. International Organization for Standardization (ISO). Downloaded 2026, January 26, from https://www.iso.org/standard/42001

Ivković, R. (2025). Transformation and economic aspects of software engineering through the implementation of the EU AI ACT. Pravo - teorija i praksa, 42(1), pp. 186–200. DOI: 10.5937/ptp2501186I

Krause, A., & Hübotter, J. (2025). Probabilistic Artificial Intelligence. arXiv preprint arXiv:2502.05244v1. DOI: 10.48550/arXiv.2502.05244

Meyer, B., & Weber, R. (2025). Programming Really Is Simple Mathematics. arXiv preprint arXiv:2502.17149. DOI: 10.48550/arXiv.2502.17149

Mladenov, M. (2023). Human vs. artificial intelligence – EU’s legal response. Pravo - teorija i praksa, 40(suppl.), pp. 32–43. DOI: 10.5937/ptp2300032M

Moine, A., Westrick, S., & Tassarotti, J. (2025). All for One and One for All: Program Logics for Exploiting Internal Determinism in Parallel Programs. arXiv preprint arXiv:2511.23283. DOI: 10.48550/arXiv.2511.23283

National Institute of Standards and Technology (NIST). (2023). AI Risk Management Framework (AI RMF 1.0). NIST. Downloaded 2026, January 22, from https://www.nist. gov/itl/ai-risk-management-framework

Quinon, P. (2025). Intensional Differences Between Programming Languages: A Conceptual and Practical Analysis. Philosophies, 10(6), p. 129. DOI: 10.3390/philosophies10060129

Regulates on artificial intelligence (Artificial Intelligence Act - 2024/1689). Official Journal of the European Union (OJ). Downloaded 2026, January 25, from https://eur-lex.europa.eu/eli/reg/2024/1689/oj

Wachter, S., Mittelstadt, B., & Russell, C. (2021). Why fairness cannot be automated: Bridging the gap between EU non-discrimination law and AI. Computer Law & Security Review, 41, 105567. DOI: 10.1016/j.clsr.2021.105567

Zakon o obligacionim odnosima. [Law on Obligations]. Službeni list SFRJ, br. 29/78, 39/85, 45/89 ‒ odluka USJ i 57/89, Službeni list SRJ, br. 31/93, Službeni list SCG, br. 1/03 ‒ Ustavna povelja i Službeni glasnik RS, br. 18/20

Zakon o zaštiti potrošača [Law on Consumer Protection]. Službeni glasnik RS, br. 35/26