examining the potential challenges | 10.1163/22131035-14020001 17 Since nationality laws can change, country-related information is continuously updated so that changes in nationality law and practice in relevant countries are taken into account. Just like human decision-makers ai will also require regular updates.85 ai algorithms can autonomously update themselves. Continuous or online learning is a type of machine learning technique where a model learns from new data streams without being re-trained.86 This type of learning adapts instantly to keep up with changes and trends in real time.87 In a situation where an ai system does not have access to real-time data, ai will not pick up on any recent repeals, amendments or publications of new legislation. ai algorithms may also generate information that are inaccurate.88 Therefore, accuracy may be compromised in both instances.89 Humans remain more reliable, as they can actively seek information from relevant authorities to ensure accuracy and avoid relying on outdated or incorrect data. For country-related training data, which an ai must learn, it has to be treated as accurate and needs to be obtained from reliable and unbiased sources, preferably more than one.90 Therefore, information sourced from State bodies directly involved in nationality mechanisms in the relevant State, or non-State actors which have built up expertise in monitoring or reviewing such matters, is preferred.91 They can also be acquired from human rights organizations, non-governmental organizations (ngo’s) dealing with nationality issues, and international bodies (e.g., unhcr, International Organization for Migration (iom), Institute Statelessness and Inclusion (isi)) including Independent researchers such as Citizenship Rights Initiative in Africa. International organisations and researchers might not have up to date or accurate information on states. Even State bodies directly involved in nationality mechanisms in the relevant State may lack accurate data capturing the extent of the issue of statelessness.92 According to unhcr, the true number 85 86 87 88 89 90 91 92 Ibid, para. 86. Cecilia S Lee and Aaron Y Lee, ‘Applications of Continual Learning Machine Learning in Clinical Practice’ (2020) 2(6) The Lancet Digital Health 1. Ibid. Mavundla v mec: Department of Co-Operative Government and Traditional Affairs KwaZuluNatal and Others (7940/2024P) [2025] zakzphc 2; 2025 (3) sa 534 (kzp) (8 January 2025). See Clyde & Co, ���Regulation of the Use of Artificial Intelligence in Law Firms’, <https://www.clydeco.com/en/insights/2025/04/regulation-of-the-use-of-artificial -intelligence> accessed 30 September 2025. Handbook on Statelessness, (n 13) para. 86. Ibid, para. 85. Adeyemi Saheed Badewa,‘Statelessness, development, and protection of ‘disadvantaged groups’: Bridging the post-2030 sustainable development gaps’ (2022) 8 (3) African Human Mobilty Review 58–59. International Human Rights Law Review (2025) 1–31