examining the potential challenges | 10.1163/22131035-14020001 13 in decision-making.61 However, Liodden asserts that consistency does not necessarily guarantee accuracy, as decisions can be consistent yet not accurate.62 ai, however, has a reputation for exacerbating bias because it lacks the ability to distinguish bias from unbiased data set.63 ai lacks the ability to understand the underlying moral reasoning behind ethical decisions. Although humans possess emotional intelligence, moral reasoning and experience, they are still susceptible to bias. The key difference is that ai lacks moral awareness and cannot take responsibility for its decisions. While humans can be influenced by others, they retain the ability to choose and to refrain from acting. In contrast, ai merely replicates and reinforces the patterns it has been trained on, without the capacity for independent judgment due to the fact that these algorithms are highly dependent on the data they are trained on.64 Decisions are based on data; if the data is biased, the output will likely reinforce inevitable bias. Bias tends to affect groups that have been historically discriminated and marginalised based on gender, social class, or race, but not in all cases.65 For instance, in 2014, Amazon’s ai recruiting tool which was used to review resumes of job applicants for software developer positions and other technical roles, discriminated against women by favouring men over women because it was trained on the profiles of male successful Amazon employees.66 It was fed resumé data from the existing pool of Amazon software engineers, 61 62 63 64 65 66 Stephen Legomsky, ‘Learning to Live with Unequal Justice: Asylum and the Limits to Consistency (2007) 60 Stanford Law Review 426. Tone Liodden, ‘Making the Right Decision: Justice in the Asylum Bureaucracy in Norway’ in Nick Gill and Anthony Good (Eds.), Asylum Determination in Europe: Ethnographic Perspectives (Springer International Publishing 2019) 245. Jungmin Choi and Melody Chao, ‘For Me or Against Me? Reactions to ai (vs. Human) Decisions that are Favorable or Unfavorable to the Self and the Role of Fairness Perception’ (2024) Personality and Social Psychology Bulletin 2. fra European Union Agency for Fundamental Rights, ‘Bias in Algorithms – Artificial Intelligence and Discrimination’ Vienna, 2022 accessed 3 October 2025. ‘What is ai? Artificial Intelligence Explained’ <https://www.techtarget.com /searchenterpriseai/definition/AI-Artificial-Intelligence> accessed 30 September 2025. Lorenzo Belenguer, ‘ai Bias: Exploring Discriminatory Algorithmic Decision-making Models and the Application of Possible Machine-centric Solutions adapted from the Pharmaceutical Industry’(2022) 2 ai and Ethics 773–774; Rowena Rodrigues, ‘Legal and Human Rights Issues of ai: Gaps, Challenges and Vulnerabilities’(2020) 4 Journal of Responsible Technology 2. Jeffrey Dastin, ‘Insight—Amazon Scraps Secret ai Recruiting Tool that Showed Bias Against Women’ 11 October 2018 <https://www.reuters.com/article/us-amazon-com-jobs -automation-insight-idUSKCN1MK08G/> accessed 30 September 2025. International Human Rights Law Review (2025) 1–31