Velomahanina Tahinjanahary Razakamaharavo
“It is frightening but it’s the reality of where war technology is taking us” (General John R. Allen)
Numbers prove it: $15,2 billion funding invested in AI startups in 2017 and $103.1 billion worth of spending in robotics and drones in 2018. These are among the statistics that have been reported on the status of Artificial Intelligence in various sectors. These demonstrate how far we have come since 1950 when Alan Turing published his paper and the term’ Artificial intelligence’ (AI) was coined for the first time by John McCarthy. A report forecasting AI in the Military Market by offering clearly says that there is a projected growth from $6.26 Billion in 2017 to $18.82 Billion by 2025. The exponential development of AI in the private sector shows how swiftly this technology is evolving. Currently, the public sector is also starting to follow suit although it is still largely left behind. Moreover, it is clear now that many nation-states are fully engaged in the so-called AI race. But how is AI impacting organisations working on peace operations? Can one say AI is reshaping these operations?
Whenever it comes to talking about peace operations, the two first organisations that come to mind are the United Nations with its peacekeeping operations (UNPKOs) and the North Atlantic Treaty Organization (NATO). Since the nature of the work of these two organisations is in terms of preserving world peace, it is first necessary to discuss what AI applications within the framework of their daily operations can one talk about. AI is what I consider to be a disruptive and an enabling technology. It is a technology that is diverse and has infinite possibilities of development and therefore its many applications could cover various domains/tasks in peace operations and warfare in general. Although it has been very common to hear about smart-peacekeeping, tech-enabled UN operations and smart NATO, AI is approached and integrated differently within these two. NATO is more advanced and at the UN, one can say there is an emerging interest.
First, within the UN, on the whole AI is this big uncertain thing that many fear a lot while very few decide to embrace it and see how to live with it. Some actors, such as the World Food Program, are very optimistic regarding its use. What is fascinating is that now we see more and more interest in regulating the technology. Thus, Mexico launched general debates on AI’s regulations and its future. Moreover, there is already a Group of Governmental Experts on lethal autonomous weapons systems (LAWS) which is said to have the backing of around 25% of the UN member states. The UN also set up an agency monitoring threats from AI and robotics. For now, within the Department of Peace Operations/ DPO (and that includes the UN Security Council, the C34 committee etc.) there is no specific ongoing discussion on AI applications in Peacekeeping Operations (PKOs). That being said, AI was discussed during the 4th International Partnership for Technology in Peacekeeping: “Next Tech Peacekeeping” symposium where it was considered to be a future peacekeeping tech.
Second, one can say that NATO might be currently undergoing significant changes in terms of smartening peace operations as, during the last two years, it has been proactively engaging with AI in various fields. NATO is testing diverse applications of AI. There are certain patterns one can notice regarding where the priorities are. For instance, it invests a lot in intelligence, surveillance, and reconnaissance (ISR) activities/ situational awareness. This explains the strong narrative on data collection, processing and analysis for decision-making where priority has been put on big data, cloud computing, quantum computing, and cyber-defence/ information warfare. And still in the domain of decision-making, another area where NATO spends a lot of resources is autonomy. It prioritizes human-machine teaming exploring, for instance, the possibility of using (semi-)autonomous robots. In 2018, NATO’s biggest exercise “trident juncture involved more than 20 AI-related experiments, including an effort to autonomously detect, diagnose, and deliver care to wounded soldiers suggesting for a more rapid process compared to the care provided only by humans”. Other tests also involve the development of high-tech equipment, for instance, exploring the possibility to 3D-print parts for weapons and deliver them by drone.
For now, NATO relies a lot on strategic foresight, analysis, and testing whenever it comes to talking about AI applications. However, it is moving fast and the Allied Command Transformation (ACT) in North-Folk Virginia is taking charge. Within that framework, it has changed its approach and strategy which is working usually with the traditional defence industry by including academia and the private sector, especially start-ups in the process. It also says that it puts civilians at the centre of the process. For example, in late 2018, NATO-ACT launched a call for solutions (“Request for Information”) based on new and future technologies (including AI) supporting the improvement of its Lessons Learned Process. The participants were selected to present their solutions to the NATO chiefs and among them are for instance Amazon, Indexima, Speech Processing Solution-Philips, and Lockheed Martin Advanced Technology Laboratories. Indeed, NATO is moving forward. The former head of ACT General Mercier was developing a proof of concept to get adherence of AI-skeptical NATO members as the alliance he argued was slower to adopt these technologies and integrate them in daily operations when compared with powerful nation-states.
All of these demonstrate the willingness to “AI-smarten” peace operations. However, DPO and NATO, again, due to the nature of the operations they are running, the structure of their forces, the inter-governmental processes, the procedures and the contributing countries involved in force generation etc. do not have the same approach to AI. UN peacekeeping operations are very slow to react and adapt to new technologies, the UN probably won’t be able to afford what NATO is currently doing to integrate Artificial intelligence in the operations. For a specific policy, and it depends on the type of action involved or concept to be integrated, it would take from two to ten years for it to be fully implemented within these organizations. It is even argued, at NATO for instance, for some concepts “on average it takes 16 years from conception of military capability to operational effect, which is far too long”. Within the UN Department of Peacekeeping Operations (DPKO) and Department of Field Support (DFS) for example, these would go through inter-governmental processes, workshops with various stakeholders and even ‘brownbag’ lunches, long and strenuous negotiations over languages among member states, implementation of various activities and projects, without forgetting the never-ending “begging for money” for sponsorship.
That being said, by having a look at what has been done so far at the level of the UN in general, or among other forums where peacekeeping operations are discussed, and within NATO, AI has great potential. It is even reshaping peace operations, especially because member states within their national defence forces and NATO, as previously explained, already have AI projects, or some of its members and partner countries have national strategies (e.g.: US’ project Maven, China, India etc.). Fear and excitement co-exist in the narratives on AI as its use can be both kinetic and non-lethal. Applications of AI in peace operations would require both DPO and NATO to engage beyond discussing and planning policies as well as strategies at the tactical and operational levels, such as integrating AI in doctrines or testing. When examining what has been done so far, many pieces of the puzzle are still missing. For instance, it is mandatory to make a significant exploration of the ethical aspects of all types of AI applications in peace operations (not only issues concerning Lethal Autonomous Weapon Systems which are still gray zones and I must admit the NATO discussions at this level are not really open to the general public). One must recognize that indeed NATO has started to adopt a stance where it prioritizes the engagement of civilians in the process, which is a sign of a nascent democratic governance of AI, but still, for now, it fails to include the grassroots in its discussions. Efforts should also be made at the local/ political and supra-national level to get the buy-in of all contributing countries involved in peace operations, accepting that AI is becoming more and more ubiquitous. Thus, one way or another it will be required to weigh its implications and probably integrate it into the whole process. It is necessary to push discussions further both at the UN and NATO and introduce regulations. Finally, it is high time to start focusing already on building capabilities, investing in change management and training.
Velomahanina Tahinjanahary Razakamaharavo is a Scientific Collaborator at UCLouvain in Belgium. She works on gender in UN and NATO peace operations as well as the implications of Artificial Intelligence in the fields of peace and security. The views expressed in this paper are solely the author’s own.
Image Credit: Pixabay