The Role of the African Union Peace and Security Council in Promoting the Responsible use of Artificial Intelligence for Refugees and Asylum Seekers in Africa

Petra Molnar wrote that the development of technology ‘reinforces power asymmetries between countries’ and influences ‘our thinking around which countries can push for innovation, while other spaces like conflict zones and refugee camps become sites of experimentation.’¹Molnar, P.; European Digital Rights (EDRi); and the Refugee Law Lab (2020) ‘Technological Testing Grounds: Migration Management Experiments and Reflections from the Ground Up’, Available at: https://edri.org/wp-content/uploads/2020/11/Technological-Testing-Grounds.pdf (Date accessed: 24 August 2024), p. 16. Africa has been dealing with a perennial refugee crisis. According to the United Nations High Commissioner for Refugees (UNHCR), ‘Around 30 million internally displaced persons, refugees and asylum-seekers live in Africa, representing almost one-third of the world’s refugee population.’²UNHCR (2020) ‘Where we work: Africa’, Available at: https://www.unhcr.org/about-unhcr/where-we-work/africa (Date accessed: 24 August 2024). As artificial intelligence (AI) is gradually reshaping the way things are done all over the world, governments are turning to novel technologies to streamline processes and reduce administrative burdens and costs in public service delivery, including critical decisions affecting the rights of refugees and asylum seekers. While the deployment of AI for border control seems to be more pronounced in developed countries, it is gradually gaining traction in Africa. African governments are exploring the best ways to harness AI, not just for economic growth but also for improving their security architecture and border control. Indeed, AI can be of immense assistance to Africa in managing the refugee crisis. However, there is a need for deeper engagement on how these technologies may compromise the rights of refugees and asylum seekers in Africa, as it is being done in developed countries today. Whereas a harmful use of AI in refugee settings in Africa can exacerbate the existing crisis, a responsible use of AI can have alleviatory impacts. This research examines the role of the African Union (AU) Peace and Security Council (PSC) in this regard.

Artificial Intelligence and Refugee Control

There is no generally accepted definition of AI as yet, but it has been described by the UNHCR as:

a science and set of computational technologies, that are inspired, but typically operate quite differently from, the way people use their nervous systems and bodies to sense, learn, reason, and take action.³Stone, P; Brooks, R.; Brynjolfsson, E.; Calo, R.; Etzioni, O.; Hager, G. et al. (2016) ‘Artificial Intelligence and Life in 2030: One Hundred Year Study on Artificial Intelligence’, Report of the 2015 Study Panel, Stanford University, Available at: https://ai100.stanford.edu/2016-report (Date accessed: 30 August 2024), p. 4.

AI technologies refer to ‘systems or applications that have the ability to interpret and learn from data for the performance of identified tasks in an agile way.’⁴Eke, D. O.; Chintu, S. S.; and Wakunuma, K. (2023) ‘Towards Shaping the Future of Responsible AI in Africa’, In: Eke, D. O.; Wakunuma, K.; and Akintoye, S. (Eds), Responsible AI in Africa Challenges and Opportunities, Cham, Switzerland: Springer Nature (pp. 169–194), p. 172. AI is being deployed in developed countries ‘for the purpose of detecting, preventing and combating illegal immigration and cross-border crime and contributing to ensuring the protection and saving the lives of migrants.’⁵Marino, S. (2021) Mediating the Refugee Crisis Digital Solidarity, Humanitarian Technologies and Border Regimes, London: Palgrave Macmillan, p. 51. It is apposite to note that there are advantages of AI in refugee contexts even for Africa. For example, AI GPS navigation systems can assist refugees in navigating unfamiliar terrains. AI can make cumbersome processes more efficient for border controllers, expedite data collection and analytics, and help governments be more responsive during the earlier stages of crisis situations through climate or crisis prediction tools. However, there are certain areas in which the deployment of AI tools in refugee settings can have negative consequences for refugees and asylum seekers. These include the following:

1. AI drones and infrared cameras used to detect movement near or around physical borders. 
2. AI biometric technologies, such as fingerprint and iris scanners and facial recognition technologies for identity verification. 
3. Dialect detectors to confirm the asylum seekers’ dialect claims. 
4. Lie detectors to gauge the credibility of asylum claims. 
5. ‘Robodogs’ to keep ‘illegal’ migrants off the borders. 
6. AI-enabled micro-gesture and emotion analysis tools to identify or ascertain the vulnerability status of asylum seekers.⁶Forster, M. (2022) ‘Refugee Protection in the Artificial Intelligence Era: A Test Case for Rights’, Research paper, Royal Institute of International Affairs, p. 6.

Risks Involved in Deploying AI in Refugee Settings in Africa

Deploying AI technologies on refugees can have far-reaching effects for their rights to asylum, privacy and non-discrimination, as enshrined in the African Charter on Human and Peoples’ Rights of 1981 (Banjul Charter), as well as the principle of non-refoulement entrenched in the Organisation of African Unity’s (OAU) 1969 Convention Governing the Specific Aspects of Refugee Problems in Africa (OAU Refugee Convention). For instance, excessive border surveillance on asylum seekers and refugees can force them to resort to dangerous alternatives that may include returning to a territory where their lives, physical integrity or liberty would be threatened. This violates the principle of non-refoulement enshrined in Article 2(3) of the OAU Refugee Convention.⁷AU (1974) ‘OAU Convention Governing the Specific Aspects of Refugee Problems in Africa’, Available at: https://au.int/sites/default/files/treaties/36400-treaty-36400-treaty-oau_convention_1963.pdf (Date accessed: 30 August 2024), p. 7.

Facial recognition tools pose the risk of data protection breaches and can exacerbate gender and other kinds of discrimination against refugees. The use of lie or emotion detectors on refugees disrespects the dignity of their human person and neglects the ordeal they have endured. Forced from their homes due to violent conflict, environmental disasters or ethno-religious persecution, these individuals, especially women and children, embark on a dangerous journey to save their lives. As noted by Okeke, ‘the stories of African migrants and refugees are gory and shameful in our acclaimed civilised world.’⁸Okeke, K. N. C. (2021) ‘Posttraumatic Stress Disorder (PTSD) in African Migrant/Refugee Population’, In: Abidde, S. O. (Ed.), The Challenges of Refugees and Internally Displaced Persons in Africa, Cham, Switzerland: Springer Nature (pp. 175–198), p. 176. On a continent where most countries are still struggling to cater for their own citizens, the fate of refugees hangs in the air. Deploying high-risk AI technologies on them may be problematic. As the deployment of AI border technologies is still nascent, African leaders have the opportunity to proactively set the moral tone for the use of such technologies in refugee settings in Africa. In this regard, there is a need for engagements on preventive policy measures that can best safeguard the rights of refugees in Africa from the adverse effects of high-risk AI. Article 2(1) of the OAU Refugee Convention provides that:

Member States shall use their best endeavours consistent with their respective legislations to receive refugees and to secure the settlement of those refugees who, for well-founded reasons, are unable or unwilling to return to their country of origin or nationality.⁹AU (1974) ‘OAU Convention Governing the Specific Aspects of Refugee Problems in Africa’, op. cit., p. 7.

The expression ‘best endeavours’ connotes genuine commitment on the part of states while ‘consistent with their respective legislations’ clearly excludes the use of dangerous AI technologies that undermine the rights of refugees under the law. This provision paves the way for a rights-based and Afrocentric approach to the use of AI for refugee management in Africa.

The Role of the PSC in Promoting the Responsible use of AI in African Refugee Settings

The PSC was established by the Protocol Relating to the Establishment of the Peace and Security Council adopted in 2002 (the Protocol) pursuant to Article 5(2) of the AU Constitutive Act. Article 2(1) of the Protocol provides that the PSC shall be:

a standing decision-making organ for the prevention, management and resolution of conflicts and a security and early-warning arrangement to facilitate timely and efficient response to conflict and conflict situations in Africa.¹⁰AU (2002) ‘Protocol relating to the establishment of the Peace and Security Council of the African Union’, Available at: https://www.peaceau.org/uploads/psc-protocol-en.pdf (Date accessed: 30 August 2024), p. 5.

The PSC’s functions include the promotion of peace, security and stability in Africa, early warning and diplomacy, humanitarian action, and disaster management. The Protocol also empowers the PSC to ‘anticipate and prevent disputes and conflicts, as well as policies that may lead to genocide and crimes against humanity’; ‘support and facilitate humanitarian action in situations of armed conflicts or major natural disasters’; ‘protect human rights and fundamental freedoms, respect for the sanctity of human life and international humanitarian law’; and ‘promote and encourage the implementation of […] Conventions and Treaties on arms control and disarmament.’¹¹Ibid, pp. 5, 9, 10.

From the foregoing, the role of the PSC in regulating the use of AI technologies in refugee settings includes monitoring and ensuring that the use of AI technologies by African states in border control does not violate human rights and humanitarian law. In this connection, the PSC can investigate any host state that deploys high-risk AI technologies on refugees. Other roles include monitoring and discouraging AI policies that can lead to crimes against humanity, such as AI Robodogs, and timeously detecting and preventing the triggers of refugee crises, such as armed conflict and natural disasters in Africa. To achieve these roles, the Panel of the Wise and the Continental Early-Warning System (CEWS), established under Article 11 and 12 of the Protocol, must be revived and equipped. This is because the PSC will depend on advice from these subsidiary bodies for monitoring and data collection and analysis on simmering, potential, actual and post-conflict initiatives and activities in Africa related to refugees. This would necessitate the strengthening of the continent’s Situation Room to function better. The Situation Room is an observation and monitoring centre at the Conflict Management Directorate of the AU. It forms part of the CEWS and is ‘responsible for data collection and analysis on the basis of an appropriate early warning indicators module.’¹²AU PSC (2006) ‘Report on the status of the establishment of the Continental Peace and Security Architecture’, 57th meeting of the AU PSC, 21 June, Addis Ababa, Ethiopia, Available at: https://www.peaceau.org/uploads/psc-protocol-en.pdf (Date accessed: 30 August 2024), p. 3.

The PSC can leverage AI for the effective implementation of its early-warning role. At its 1214th meeting held on 13 June 2024 in Addis Ababa, Ethiopia, the PSC thus requested the AU Commission to promote the use of AI-driven early-warning systems to detect and analyse potential conflict indicators, enabling proactive conflict prevention measures. Being able to predict forced displacement can prepare humanitarian aid agencies in advance to provide refugees with shelter, food and medical care. Some African countries are already beginning to utilise this AI for these purposes. For example, in Somalia, AI has been used for predictive analytics to address the problem of forced displacement proactively.¹³Hoffmann Pham, K. and Luengo-Oroz, M. (2022) ‘Predictive modeling of movements of refugees and internally displaced people: Towards a computational framework’, Journal of Ethnic and Migration Studies, 49(2), pp. 408–444, Available at: https://arxiv.org/pdf/2201.08006 (Date accessed: 30 August 2024), p. 18. In addition, Uganda has an AI-powered radio content analysis tool that analyses radio discussions in different parts of the country about topical issues, including refugees. The aim is to gauge the sentiment of local communities and refugees. According to Microsoft, this tool ‘aids in the strategic planning and implementation of various programmes aimed at easing tensions and fostering better integration between refugees and host communities.’¹⁴Microsoft (2024) ‘AI in Africa: Meeting the Opportunity’, Available at: https://blogs.microsoft.com/wp-content/uploads/prod/sites/5/2024/01/AI-in-Africa-Meeting-the-Opportunity.pdf (Date accessed: 26 August 2024), p. 63. This is in line with the principles of humanitarianism and solidarity underpinned by the OAU Refugee Convention as the right approach to refugee crises in Africa. The PSC and other African countries can emulate Uganda in leveraging AI as an early-warning system related to diverse refugee issues, including climate prediction and analytics for disaster management.

Challenges and Recommendations

The first challenge is the possible impact of AI on refugees and asylum seekers in Africa. As noted earlier in this paper, the PSC has urged the AU Commission to promote the use of AI as a proactive conflict prevention measure. This is a step in the right direction. However, some scholars have noted that African AI policy discourse is still embryonic.¹⁵Gwagwa, A.; Kraemer-Mbula, E.; Rizk, N.; Rutenberg, I.; and de Beer, J. (2020) ‘Artificial Intelligence (AI) Deployments in Africa: Benefits, Challenges and Policy Dimensions’, African Journal of Information and Communication, 26, 1–28, p. 4. This is even more so for policy discourses on AI and refugees. Similarly, galvanising a positive attitude and political will on the part of states towards refugees and asylum seekers in Africa requires much work on the part of the PSC. Zard, Beyani and Odinkalu believe that the crisis facing refugees on the continent reflects not a paucity of norms, but rather a failure to implement them.¹⁶Zard, M.; Beyani, C.; and Odinkalu, C. A. (2003) ‘Refugees and the African Commission on Human and Peoples’ Rights’, Forced Migration Review, 16, 33-35, Available at: https://www.fmreview.org/zard-beyani-anselmodinkalu (Date accessed: 28 August 2024). This implementation crisis may be exacerbated in the AI era if appropriate steps are not taken.

Finally, the lack of AI knowledge and capacity is also a challenge in Africa. However, this can be overcome through intentional AI capacity building and investment, not just for researchers but also for leaders. A collective approach to AI learning and regulation will require members of the Panel of the Wise to acquire knowledge about how AI technologies work. There is also a need for improvement in the technological capacity of the Situation Room.

In conclusion, the PSC has a vital role to play in the promotion of responsible use of AI technologies for refugees and asylum seekers in Africa. Through its conflict monitoring, crisis detection and early-warning systems, it can work closely with African leaders to deal with refugee problems proactively and in a manner that respects the Banjul Charter and the OAU Refugee Convention. To achieve its role, the PSC, in line with Article 17 of the Protocol, can cooperate with the UN Security Council and the UNHCR and, where necessary, request financial assistance from the UN. In pursuing its role as discussed above, the PSC may face other practical bottlenecks. However, with constant effort and genuine commitment, the PSC can utilise its critical position on the continent to drive the right approach in the use of AI in Africa’s refugee settings. As rightly stated by Madeleine Forster, ‘how we treat asylum seekers and refugees interacting with AI will be a test case for emerging domestic and regional legislation and governance of AI.’¹⁷Forster, M. (2020) ‘Refugee Protection in the Artificial Intelligence Era’, op. cit., p. 1.