8 10.1163/22131035-14020001 | enigbokan being explicitly programmed for specific tasks (deep learning neural networks and artificial neural networks a subset of machine learning).33 This differs from the traditional approach to ai which involved a programmer trying to translate the way humans make decisions into software code.34 In other words, the programmer does not need to write all the instructions that the system should carry out for it to perform a task. Instead, the system is fed with training data in the form of examples and utilises the generalised algorithm to analyse real data, discern patterns and produce an output.35 The data set used is what powers the algorithm in the decision-making process. During machine learning, the algorithms are fed with historic data, usually broken into training data and test data to validate the results.36 The training data would be used to design and operationalize the system, while the test data is used to access the predictive accuracy of the algorithms. This allows the human analysts to examine the predictive accuracy of the algorithm before real-life data is introduced into the model (the programme that has been trained) to make decisions.37 Supervised algorithms are a type of machine learning given a labelled outcome variable (also known as an output or response variable) representing the true values to be predicted on the basis of input training data.38 In contrast, 33 34 35 36 37 38 ibm, ‘What is Artificial Intelligence (ai)?’<https://www.ibm.com/topics/artificial -intelligence> accessed 30 October 2025; Amazon, ‘What is a Neural Network?’ <https://aws .amazon.com/what-is/neural-network/> accessed 30 October 2025: Deep learning ‘is an advanced form of machine learning that allows software to ‘train itself to perform tasks, like speech and image recognition, by exposing multilayered neural networks to vast amounts of data.’ Though in this case features are not extracted by humans. Rather, data sets are fed directly into the deep learning algorithm, which then predicts the occurrence of objects. Deep learning algorithms do this via multiple layers of artificial neural networks that mimic biological brains. Access Now, ‘Human Rights in the Age of Artificial Intelligence’ 8 <https://www .accessnow.org/wp-content/uploads/2018/11/AI-and-Human-Rights.pdf> accessed 30 September 2025; ibm ‘ai vs. Machine Learning vs. Deep Learning vs. Neural Networks: What’s the Difference?’ <https://www.ibm.com/think/topics/ai-vs-machine-learning-vs -deep-learning-vs-neural-networks> accessed 30 September 2025; Jones, (n 32) at 5. Christian Janiesch, Patrick Zschech, and Kia Heinrich, ‘Machine Leaning and Deep Learning’ (2021) 31 The International Journal on Networked Business 686. Awotula, (n 2) at 6. Yavar Bathaee,‘The Artificial Intelligence Black Box and the Failure of Intent and Causation’ 2018 31(2) Harvard Journal of Law & Technology 900; Awotula, (n 2) at 6. David Lehr and Paul Ohm, ‘Playing with the Data: What Legal Scholars Should Learn about Machine Learning’ (2017) 51 uc Davis Law Review 673; Janiesch and others, (n 35) at 686; Lucia Nalbandian ‘Using Machine-Learning to Triage Canada’s Temporary Resident Visa Applications’ Working Paper No. 2021/9 at 2–3. International Human Rights Law Review (2025) 1–31