Over the last seventy years, Artificial Intelligence (AI) has made incredible progress in leaps and bounds. First introduced by John McCarthy during the 1950s, he described AI as the science and engineering of making intelligent machines, especially intelligent computers and programmes studied (McCarthy, 2007). More specifically, he invented the LISP programming language in 1958 using the lambda calculus, which was a major milestone in the development of advanced artificial intelligence applications (Allganize, 2020). Thereafter, AI has expanded its ties in several fields, including the civilian and military sectors. On the one hand, big companies such as Amazon and Google have used these tools to build vast commercial empires based in part on predicting the wants and needs of the people who use them (Gatopoulos, Aljazeera, 2021). On the other hand, the development of AI in the military originated a few decades before and has been more intense and challenging.
Between the 1960s and the 1970s, the US had already used AI to gather different pieces of information in Vietnam, including the transit of soldiers and trucks, by using microphones and sensors. Once data was collected, this information was sent to drones and computers. A few minutes later, warplanes started carpet bombing a piece of terrain determined by an algorithm. This was the Igloo White mission, which took place in 1970. Since then, the importance of technology has progressively increased, becoming the key driver of the military. Nowadays, it is largely acknowledged that the superiority of Artificial Intelligence represents the new paradigm of power between superpowers (Kumar Jha and Das, BusinessWorld, 2021). The Russian President Vladimir Putin confirmed this new trend in 2017 when he stated that “Whoever becomes the leader in this sphere will become the ruler of the world” (Putin, RT, 2017).
So far, much progress has been achieved in the air and sea domains. The US has employed Unmanned Aerial Vehicles (UAVs) in several military missions, especially to fight the Taliban in Afghanistan. That is the case of MQ-9 ‘Reaper’, which has been primarily employed as an intelligence-collection asset, and, secondarily, against dynamic execution targets (U.S. Air Force, 2021). In addition, its features, such as wide-range sensors, precision weapons, and loiter time, provide a unique capability to perform strikes and reconnaissance against high-value and time-sensitive targets (U.S. Air Force, 2021). Moreover, the US has tested a naval drone called the Sea Hunter, an Unmanned Surface Vessel (USV), able to navigate autonomously for three months. The prototype has the potential to traverse thousands of kilometres of open ocean for months without a single crew member aboard and at a fraction of current costs, paving the way for a whole new class of ocean-going vessels (Naval Technology, 2018). The same applies to France, which is working with Thales on the production of AUSS drones, able to operate both underwater and on the surface. With its 360° agility and superior endurance, the AUSS system can perform many civil and military missions that were formerly considered impossible.
When it comes to AI’s impact on soldiers, the situation seems quite different. Due to its complexity, the land domain will be the last one affected by these strategic changes. However, it is worth noting that a step in that direction was undertaken by the Defense Advanced Research Projects Agency (DARPA), which led the Talos programme to produce a robotic exoskeleton. The programme aimed at developing a special uniform able to strengthen soldiers’ resistance, but it also was intended to include several other features, such as a liquid that freezes in the event of impact and a mechanism to help clot blood in the event of bleeding. Moreover, it should have included a plaster to prevent dehydration, contact lenses for night vision, and smart ammunition – programmed to explode after a certain flight time. However, although the Talos programme started in 2013, it came to an end in February 2019, when James Smith, SOCOM’s acquisition executive, stated that the final product would not match the sales pitch (Tucker, Defense One, 2019).
The failure of the Talos programme says a lot about the current relation between AI and soldiers, as there are several obstacles along the way. From an ethical point of view, implications of AI have largely been debated since it lacks common sense and a human way of thinking. Despite all efforts to improve AI, for example, it will never fully understand the consequences of its actions. This leads to other questions, namely the inability of these systems to explain how they made their decisions and their use in war scenarios. In other words, most of what occurs inside an AI system is a black box, and there is very little that a human can do to understand how the system makes its decision (Maxwell, Modern War Institute, 2020). Moreover, due to the inability to understand the consequences of their actions, AI systems will hardly be able to tell friend from foe or simply fighters from civilians. Nevertheless, the issues mentioned above are just the tip of the iceberg. Another crucial point regards the inability of AI systems to multi-task as well as to understand inputs and context within the inputs. Unlike humans, they can neither distinguish specific enemy vehicles nor decide which weapon could be used against them. Again, they are also very poor at understanding inputs and their context: AI does not understand what the image is, it simply learns textures and gradients of the image’s pixels (Maxwell, 2020).
It is also worth noting AI’s achievements so far. Over the last few years, AI has improved significantly on several tasks, such as image recognition, recommendation systems, and language translation. Among these categories, image recognition has proven to be one of the most successful applications of AI due to the system’s ability to scan images, follow multiple potential targets, and collect and analyse large amounts of data rapidly (Gatopoulos, 2021). Therefore, it could be posited that the most successful applications of AI can be observed when there are large quantities of labelled data, like ImageNet, Google Translate, and text generation (Maxwell, 2020).
Although AI has made great strides in the military over the last four decades, it is worth noting that not all military sectors have experienced the same level of progress. For instance, UAVs and USVs have largely improved their capabilities and features, which is confirmed by the new prototypes that continue to be released regularly. However, the same cannot be said about the relation between soldiers and AI. Efforts and trials have been made, but it appears that it is not so simple to replace the human way of thinking and decision-making, especially in a crucial role such as that of soldiers. Current outcomes seem to confirm this trend, namely the fact that AI will continue to be a tool in soldiers’ hands rather than an autonomous decision-maker or actor. At the same time, we must admit that the overall progress of AI will not stop here but will take different roads – especially in uncontested domains. Given AI’s current weaknesses in the military, it is reasonable to expect that future technological developments in the military will concern tools that are closely supervised by human experts or that have secure inputs and outputs that can provide value to the military while alleviating concerns about vulnerabilities, such as medical-imaging diagnostic tools, prediction applications, and fraud-detection programs. (Maxwell, 2020). This does not mean that AI will not impact future soldiers, but these changes will only appear in the medium to long term.
Written by Emanuele Bussagli
Allganize. (2020) How John McCarthy Shaped the Future of AI. [online] Available at: https://blog.allganize.ai/john-mccarthy/
Gagliano, G. (2020, December 12) Intelligenza artificiale, droni e robot soldato. Come sarà la nuova corsa agli armamenti. Start Magazine. https://www.startmag.it/innovazione/intelligenza-artificiale-droni-e-robot-soldato-come-sara-la-nuova-corsa-agli-armamenti/
Gatopoulos, A. (March 28, 2021) Project Force: AI and the military – a friend or foe? Aljazeera. [online] Available at: https://www.aljazeera.com/features/2021/3/28/friend-or-foe-artificial-intelligence-and-the-military
Kumar Jha, M., & Das A. (2021, August 3) AI Technology In Military Will Transform Future Warfare. BW Business World. http://www.businessworld.in/article/AI-Technology-In-Military-Will-Transform-Future-Warfare/13-08-2021-400525/
Maxwell, P. (2020, April 20) Artificial Intelligence Is The Future Of Warfare (Just Not In The Way You Think). Modern War Institute. https://mwi.usma.edu/
McCarthy, J. (2007) WHAT IS ARTIFICIAL INTELLIGENCE?. [online] Available at: http://jmc.stanford.edu/articles/whatisai/whatisai.pdf
Naval Technology. (2018). Sea Hunter: inside the US Navy’s autonomous submarine tracking vessel. [online] Available at: Sea Hunter: inside the US Navy’s autonomous submarine tracking vessel (naval-technology.com)
Putin, V. (September 1, 20171) Whoever leads in AI will rule the world. Russia Today. [online] Available at: https://www.rt.com/news/401731-ai-rule-world-putin/
Tucker, P. (February 7, 2019) The US Military Is Chopping Up Its Iron Man Suit For Parts. Defense One. [online] Available at: https://www.defenseone.com/technology/2019/02/us-military-chopping-its-iron-man-suit-parts/154706/
U.S. Air Force. (2021) MQ-9 Reaper. [online] Available at: MQ-9 Reaper > U.S. Air Force > Fact Sheet Display (af.mil)