The Application of Artificial Intelligence for Peacekeeping

The Application of Artificial Intelligence for Peacekeeping

“The increasing amount of available data, mainly due to the proliferation of access to the internet in countries where peacekeeping missions take place, has caused a technology-driven transformation of the operational environment. This comes at a time of significant developments in the fields of artificial intelligence and particularly machine learning, most of whose applications still rely on massive amounts of data. As such these developments have produced some promising individual initiatives to exploit this new and growing potential for United Nations operations.

by Hendrik Pasligh

At least as early as 1996 researchers have used machine learning (ML) to predict conflicts[1]. Today, mainly due to significantly higher amounts of available data[2], advancements in computing power and the progress made in natural language processing, several artificial intelligence (AI) tools have been added to the peacekeeping arsenal. The surge in available data has mainly been caused by the proliferation of mobile phones and access to the internet[3]. The potential for companies such as Facebook to amass profits from their access to and use of people’s data makes it lucrative for them to provide access to the internet for a much lower price[4]. Thus, the equipment is now affordable for people in relatively poorer countries, which is where most peacekeeping operations take place[5]. All of these developments constitute a technology-driven transformation of the environment of peacekeeping operations. This article seeks to explore in which ways this changing environment is characterised by an increased potential for the application of AI tools.

AI tools have the potential to support peacekeeping operations in three areas: general military tasks, conflict prediction, and specific peacekeeping tasks during an operation. Among the general military tasks is the potential for optimisation and automation of administrative processes such as logistics, increasing not the effectiveness but efficiency of military operations[6]. This could lead states to be more willing to participate in peacekeeping missions and if more personnel or equipment is deployed while costs remain constant, reduce the gap between mission goals and available resources. Additionally, AI-based virtual training can be used to enhance soldiers’ tactical abilities, thus contributing to the mission’s success while mitigating casualties[7]. Sharing such programs with less developed countries  – who overwhelmingly are the main troop contributors to United Nations (UN) missions – could compensate for the lack of resources that results in insufficient training[8] for soldiers facing complex situations in peacekeeping operations[9]. Beyond that, this would certainly be in the interest of more developed countries seeking to enable local forces to take over responsibility for the security of their own countries as with the European Union Training Mission Mali (EUTM Mali).

Several studies claim that the application of machine learning tools significantly increases success in predicting conflict[10]. This can enhance the ability to understand conflict dynamics and allow for a design of peacekeeping operations that is more appropriate for preventing re-emerging or new conflicts. For example, using ML techniques such as lasso and random forests, Blair et al. found that – contrary to previous belief – ‘the risk of local violence is higher rather than lower in communities where minority and majority ethnic groups share power’[11]. The knowledge drawn from ML-driven analysis could be used to enhance prevention capabilities[12], thereby avoiding a situation in which a peacekeeping mission has to be established in the first place. It can also be applied on a tactical level, enabling a smarter allocation of resources for day-to-day tasks[13]. An important aspect here is the use of geographic information systems (GIS) which have greatly benefited from the commercialisation and subsequent affordability of satellites and satellite imagery. Machine learning[14] tools have been utilised to analyse these massive amounts of data. Unmanned aerial vehicles (UAVs) can also contribute to this collection of data[15]. Unarmed drones were first deployed by the UN in 2013[16]. The information obtained from the use of GIS can facilitate a broad range of tasks beyond conflict prediction, including the monitoring of borders and sustainable development[17], supporting elections and other governance tasks[18] as well as field logistics[19].

Some of the most ambitious AI tools for specific peacekeeping tasks set the objective of ‘deep conflict resolution’[20], taking into account the needs and perceptions of all participants. Here, AI is used to process knowledge on conflict dynamics[21] and present it in an easily accessible way to the user in the field. This would lead to the decentralisation of knowledge and reduce a mission’s dependence on experts of psychology, conflict resolution and local culture[22]. The software cogSolv claims to provide the user with ‘options leading to truly just results’[23]. This is achieved by simulating a specific situation and, for example, suggesting to base efforts for conflict resolution on ‘Local Dignity’ instead of ‘Human Rights Discourse’.

Significant advancements in natural language processing (NLP) enabled the creation of powerful translation tools, which could enhance interoperability in multinational peacekeeping forces as well as facilitate communication with locals. Arguably even more valuable, the capability of computer programmes to process language and identify objects unlocked tools[24] to analyse open-source information, most importantly gathered from social media. A UN Global Pulse lab has used NLP to analyse radio shows in Uganda, notably including statements of people who called into the radio station, in order to gain insight into social tensions[25]. This access to unprecedented amounts and forms of information can provide peacekeeping missions with better understanding of the environment they are operating in but could also be used for smarter reactions to emergencies based on informed decision-making[26]. However, this – as well as the deployment of UAVs – raises concerns about privacy and who will have access to intelligence gathered by the UN or in the course of a UN operation[27]. Beyond that, the digitisation of UN operations might create additional vulnerabilities and establish an even more complex conflict environment by including cybersecurity risks.

In light of these numerous benefits the application of AI for peacekeeping can yield, it is worthwhile to take a look at the effort the UN is making to make the most of this technology. The UN-issued 2015 Report of the Expert Panel on Technology and Innovation in UN Peacekeeping explicitly mentions AI once – in the annexe[28]. The Secretary General’s Strategy on New Technologies recognises the importance of AI, but speaks of ‘exposure to new technologies’[29]. Nonetheless, several UN departments already make use of AI and machine learning, the Centre for Artificial Intelligence and Robotics is in the process of being established in The Hague[30] and Big Data analysis was among the main topics discussed at the 5th International Partnership for Technology in Peacekeeping Symposium in May 2019[31]. In 2013, John Karlsrud argued that in comparison to other UN activities, the possibilities of Big Data are underutilised for peacekeeping[32]. Clearly, the UN is aware of the possible benefits. However, the shift from “exploring“ and “raising awareness” to “regulating” and “committing resources” on a systemic level has yet to take place.

[1] Trappl, R / Fürnkranz, J / Petrak, J (1996) ‘Digging for Peace: Using Machine Learning Methods for Assessing International Conflict Databases,‘ paper presented at the 12th European Conference on Artificial Intelligence, Budapest, Hungary.[2] Horowitz, M C / Allen, G C / Saravalle, E / Cho, A / Frederick, K / Scharre, P (2018) Artificial Intelligence and International Security,: Center for a New American Security, p. 13.
[3] Dorn, A W (2016) Smart Peacekeeping: Toward Tech-Enabled UN Operations,: International Peace Institute, p. 1.
[4] Taylor, L / Broeders, D (2015) ‘In the name of Development: Power, profit and the datafication of the global South,’ Geoforum, Vol. 64, pp. 229-237, p. 233.
[5] World Economic Forum (2018) ‘7 charts that show how peacekeeping is changing,’ [online] available from , accessed on 3rd June 2019.
[6] Horowitz et al., p. 10-11.
[7] Cil, I / Mala, M (2010) ‘A multi-agent architecture for modelling and simulation of small military unit combat in asymmetric warfare,’ Expert Systems with Applications, Vol. 37, No. 2, pp. 1331-1343; Dormehl, L (2019) ‘The U.S. Army is building a giant VR battlefield to train soldiers virtually,’ Digital Trends, 20th March, [online] available from , accessed on 12th June 2019.
[8] McCauley, A (2014) ‘Soldiers From Poor Countries Have Become the World's Peacekeepers,’ Time, 12th September, [online] available from , accessed on 12th June 2019.
[9] Cutillo, A (2013) ‘As Peacekeeping Becomes More Complex, Progress Needed on Training,’ IPI Global Observatory, 4th September, [online] available from , accessed on 12th June 2019.
[10] Blair, R A / Blattman, C / Hartman, A (2017) ‘Predicting local violence: Evidence from a panel survey in Liberia,’ Journal of Peace Research, Vol. 54, No. 2, pp. 298-312; Perry, C (2013) ‘Machine Learning and Conflict Prediction: A Use Case,’ Stability: International Journal of Security & Development, Vol. 2, No. 3, pp. 1-18.
[11] Blair et al., p. 298.
[12] Perry, p. 2.
[13] Horowitz et al., p. 12.
[14] Naveen, J (2018) ‘4 Ways Global Defense Forces Use AI,’ Forbes, 26th August, [online] available from , accessed on 11th June 2019.
[15] Portmess, L / Romaya, B (2015) ‘Digital Peacekeepers, Drone Surveillance and Information Fusion: A Philosophical Analysis of New Peacekeeping,’ Theoria, Vol. 62, No. 145, pp. 5-22, p. 9.
[16] Dorn, p. 6.
[17] Chen, J / Li, R / Dong, W / Ge, Y / Liao, H / Cheng, Y (2015) ‘GIS-Based Borderlands Modeling and Understanding: A Perspective,’ ISPRS International Journal of Geo-Information, Vol. 4, pp. 661-676, p. 662.
[18] Convergne, E / Snyder, M R (2015) ‘Making Maps to Make Peace: Geospatial Technology as a Tool for UN Peacekeeping,’ International Peacekeeping, Vol. 22, No. 5, pp. 565-586, p. 565.
[19] Ibid., p. 567.
[20] Olsher, D J (2015) ‘New Artificial Intelligence Tools For Deep Conflict Resolution and Humanitarian Response,’ Procedia Engineering, Vol. 107, pp. 282-292.
[22] Olsher, p. 282-283.
[23] Ibid., p. 289.
[24] International Telecommunication Union (2018) United Nations Activities on Artificial Intelligence (AI), p. 32.
[25] Hidalgo-Sanchis, P (2018) ‘USING BIG DATA AND AI TO SUPPORT PEACE AND SECURITY EFFORTS IN AFRICA,’ [online] available from , accessed on 3rd June 2019.
[26] Dorn, p. 12.
[27] Portmess / Romaya, p. 8-9.
[28] Lute, J H / Bager, I J / Dorn, W / Fryer, M / Guha, A (2015) Performance Peacekeeping. Final Report of the Expert Panel on Technology and Innovation.
[29] UN Secretary General’s Strategy on New Technologies (2018), p. 13.
[30] United Nations Interregional Crime and Justice Research Institute (n.d.) ‘UNICRI Centre for Artificial Intelligence and Robotics,’ [online] available from , accessed on 3rd June 2019.
[31] UN Department of Operational Support (n.d.) ‘5th International Partnership for Technology in Peacekeeping Symposium Topics ,’ [online] available from , accessed on 7th June 2019.
[32] Karlsrud, J (2013) “Peacekeeping 4.0 Harnessing the Potential of Big Data, Social Media, and Cyber-technology,” in Jan-Frederik Kremer and Benedikt Müller, eds., Cyberspace and international relations: Theory, prospects and challenges, Springer, pp. 141-160, p. 143-144.
[33] Photo by: U.S. Air Force Graphic Illustration / Dr. Paul Hartman, 

Leave a Reply