examining the potential challenges | 10.1163/22131035-14020001 21 the human brain functions and results in decisions.113 What matters most is whether the decision can be justified with reasons grounded in fact and law. Some black-box ai systems can produce results but are unable to offer reasons or provide clear, understandable explanations for the results that were reached.114 Humans may not give reasons or clear reasons for their decisions, but they possess the ability to do so. Since the legal system demands transparency in decision-making, decision-makers can be challenged by affected parties through appeals and may be compelled to justify their decisions. As a result, humans can reflect on their decision-making processes and provide reasons.115 With ai black box models, they lack the ability to provide the rationale for their decisions even if the law imposes this duty. This is a deeper problem than just refusing to give reasons. This suggests that a full ability to provide reasons flowing from ai decision-making can be very difficult. Empirical research has shown that ai systems are unable to accurately perform the complex legal reasoning required in decision-making and as a consequence, they cannot provide legally meaningful explanations for their outcome of a case.116 Kolkman et al. asserts that ai lacks the ability to adapt its reasoning to the evolving nature of the legal system and lacks the flexibility to apply exceptions to general rules in novel cases. This is because ai algorithms tend to learn the most general rules (patterns, correlations) that appear in its training data, the algorithm cannot exercise discretion, the autonomy to decide what should be done for each individual case.117 ai algorithms is problematic because it does not follow rational evaluation or human-like reasoning. As a result, ai algorithms cannot handle complex matters that impact on individual lives.118 Explainable ai systems have been developed to provide post hoc explanations of the black-box model’s outputs. These explanations capture simplifications of a black-box system’s algorithm decision-making and do not contain all its technical details. It identifies the most important features that influenced a 113 114 115 116 117 118 John Zerilli and others, ‘Transparency in Algorithmic and Human Decision-making: Is there a Double Standard?’ (2019) 32(4) Philosophy and Technology 661–683. R v McCann [2019] uksc 34; Adamantia Rachovitsa and Niclas Johann, ‘The Human Rights Implications of the Use of ai in the Digital Welfare State: Lessons learned from the Dutch SyRI Case’ (2022) 22 Human Rights Law Review 1–15; Hassija et al, (n 10) at 46. Uwe Peters, ‘Explainable ai Lacks Regulative Reasons: Why ai and Human DecisionMaking are not Equally Opaque’ (2023) 3(3) ai and Ethics 968–969. Daan Kolkman and others, ‘Justitia ex machina: The impact of an ai system on legal decision-making and discretionary authority’ (2024) 11(2) Big Data & Society 3. Ibid. Johan Egbert Hans Korteling and others, ‘Human- Versus Artificial Intelligence’ (2021) 4 Front Artif Intell 1–10. Uwe, (n 115) at 971. International Human Rights Law Review (2025) 1–31