Artificial intelligence (AI) advances will bring about important changes not only for the economic and social sphere, but also for the military arena. An AI arms race has already begun: the US, Russia, the EU, South Korea, and China are the five biggest players currently investing in the weaponization of new technologies. While fully autonomous weapons have not been deployed yet in combat, semi-autonomous weapons have been widely used in the last decade.
The uncertainty as to how these “smart” weapons will impact decision-making during conflict makes me wonder: now that politicians can replace soldiers with tireless and emotionless robots, will they think twice before escalating conflict with another country? Will those people controlling the robot, thousands of kilometres away from their target, think twice before pushing the deadly button? Overall, will autonomous weapons lead to humans fostering more or less conflict? Can regulation aid in clearing out these uncertainties? Over the next few weeks, I will be using different case studies on the individuals behind conflict policymaking to address these questions and more.
Before I move on to the fun part, let’s first learn some background info on killer robots and International Relations theory.
What are Lethal Autonomous Weapons Systems (LAWS)?
Policymakers have not been able to reach a consensus on the definition of LAWS. The main disagreement relates to the degree of human intervention required for a weapon to be considered autonomous. The UN Convention on Conventional Weapons discussed the topic and, ironically, most (superpower) countries in favour of the development of LAWS insisted that these types of weapons had not been developed yet. Therefore, no regulation or even discussion could begin on the issue of LAWS. Developing countries disagreed, because (of course) who wants to be invaded by killer drones?
Fully autonomous weapons have not been deployed yet, but important efforts have been directed towards their development. The technology exists. These weapons would be able to make life-or-death decisions without any human intervention; humans would only be involved in their creation and programming.
However, semi-autonomous weapons have already been utilized during conflict, both by state and non-state actors, and they still give rise to the uncertainties mentioned above. After all, these weapons are remotely-controlled and, hence, don’t require human intervention on the battlefield either. Both autonomous and semi-autonomous weapons locate, select, and attack human targets without putting at risk a soldier’s life and without being subject to human emotions and exhaustion. Both weapons can target more precisely and at higher speeds, and so they are better killing tools than humans.
The main current form of semi-autonomous weapons is armed unmanned aerial vehicles (UAV), more commonly known as drones, which 68 countries already have in their possession. These remote-controlled weapons allow the operators, sitting on the other side of the globe, to watch their targets for days, and, finally, kill them. These weapons also eliminate any dangers that pilots may face in a manned vehicle during operations. A new type of war at a distance has begun.
New forms of LAWS are also being developed, such as unmanned tanks. The effectiveness and proliferation of all types of LAWS – whether autonomous or semi-autonomous – are expected to increase as technology advances and costs of production are reduced. Could global security be at risk?
Use of LAWS in Conflict
Like most security concerns nowadays, it all started with 9/11. As part of its Global War On Terror, the US was the first country that began using armed drones, specifically to carry out targeted killings of presumed terrorists in Pakistan, Afghanistan, Yemen, and Somalia. Soon other countries followed, such as France in its counter-terrorism operations in the Sahel.
Some states have used drones for targeted strikes within their domestic airspace. For instance, Turkey has used UAVs against the Kurdistan’s Workers Party. Complex intra-state conflicts that involve a multitude of parties have also experienced a rise in the use of UAVs. The Syrian war is the most prominent example: UAVs have been used, on the one hand, by Daesh and other non-state actors, and, on the other, by the US, Iran, UK, and Turkey.
Finally, UAVs have also been utilized for inter-state conflicts, leading to an increase in incidents between countries in recent years. In 2018, conflict between Israel and Iran escalated when Israeli forces downed an Iranian-made drone that had entered the country’s borders. In the Korean Peninsula, both North and South Korea have announced the use of UAVs.
UAVs have become effective weapons, thus their increased usage. Some would even say they are too effective: critics argue that LAWS have killed more civilians than terrorists or combatants. Moreover, the lack of clear regulation addressing the use of LAWS during conflict raises concerns about the effective application of international laws on war.
The upcoming blogs will examine the use of LAWS in each of the conflict types mentioned above, as well as their implications for civilians, international law, and more. Specifically, my focus will be on the people that have decision-making power during conflict and that can choose to use LAWS as a military strategy. How can the advantages and disadvantages of using LAWS affect that person’s policies when in a conflict situation?
The Policymaking Incentives at Stake
When in a conflict relation with another country or actor, a policymaker faces two main policy decisions: do I escalate this conflict and, then, for how long do I prolong it? These decisions and the people taking them are the objects of my analysis in this blog.
In order to understand policymakers’ incentives when deciding whether to escalate and prolong a conflict, the theory of offense-defense comes in handy. Let’s try to put it in simple terms:
Offense-defense theory claims that when defensive capabilities have an advantage over offensive ones, and hence defending territory is easier than taking it, war is usually avoided. The perspective that policymakers have in this regard thus becomes essential, as they are the ones making war decisions.
As for decisions on conflict escalation, when policymakers perceive an easy conquest due to their country’s offense dominance, incentives to go to war increase. This was the case in 1939 of German elites and Hitler, who over-estimated German offensive capabilities. The offense dominance of their military capabilities was real, but the elite’s perception of offense dominance was exaggerated. That is, German policymakers decided to go to war because they thought they had high probabilities of winning and that the costs of doing so would not be high (thankfully, that ended up not being the case).
Military technology is a factor that determines the perception of offense dominance. Most of us have watched movies about the highly advanced German military submarines from WWII (if you haven’t, I recommend Das Boot). Public opinion is also very influential in policymakers’ perceptions. As a general rule, the public won’t support wars because the people will be the ones that have to fight and die in them. For instance, only when countries shift from a conscription-based military to a voluntary military does public opposition to on-going wars weaken.
The use of LAWS touches upon both military technology and public opinion factors. Due to the recent acceleration in tech advances, LAWS are improving and leading to offense dominance. Although current LAWS such as armed UAVs are still not capable of avoiding attacks by air defence systems and can be easily identified and destroyed, countries are investing vast amounts of money to improve these defects. Therefore, military technology could soon shift the balance towards countries’ offense dominance by improving the effectiveness of LAWS as offensive weapons.
Moreover, the decrease in costs that LAWS entail also brings about offense dominance, which the drone expert Amy Zegart has referred to as the decrease in “blood, treasure and reputation” costs. First, LAWS entail lower financial “treasure” costs. In the case of UAVs, they are cheaper to produce and to operate than manned vehicles. For instance, the manned F-22 aircraft costs ten times as much as the unmanned MQ-9 Reaper. Deploying ground troops is also relatively expensive. Therefore, UAVs have a military financial advantage.
Secondly, life “blood” costs for the user state are also lower with the deployment of LAWS, as they do not require human intervention on the battlefield. This implication leads to the third type of reduction in costs: political “reputation” costs. The public supports the use of LAWS during conflict more than the use of human soldiers; a scenario in which conflict escalation leads to human soldiers being deployed is more susceptible to public opposition. For example, US polls show that the public “overwhelmingly favours the use of drones for counterterrorism purposes.”
While, in practical terms, LAWS will soon have an offense dominance, it does not follow suit that policymakers will actually perceive this dominance. Several studies show that, currently, there is no consensus among policymakers on whether LAWS represent defensive or offensive capabilities. Given that states are developing LAWS in secrecy and in isolation, there are divergent interpretations of the nature of LAWS.
Unfortunately, this ambiguity in the current international system does not mean that conflict escalation will be reduced. On the contrary, the uncertain, ambiguous, and unpredictable nature of LAWS could lead to either higher incentives to escalate conflict or to an unintended escalation as a result of military actions with these weapons. For instance, a policymaker from the user state may consider that the low costs of UAVs make them defensive in nature and hence less conflict escalating, whereas the policymaker from the affected state may think otherwise.
Also, the policymaker from the affected state may consider that destroying a UAV will escalate the conflict less than destroying a manned vehicle, but the policymaker from the user state may think otherwise. Ambiguity in policymakers’ perception of offense dominance can thus lead to a higher risk of conflict escalation in the short-term, at least until LAWS’ offense dominance becomes clear for policymakers. This premise is an interesting fact that I will explore in my blog posts. Could regulation help reduce the ambiguity surrounding LAWS, and hence also conflict?
Finally, regarding conflict prolongation decisions, policymakers’ incentives to extend conflict could also change as a result of LAWS. Given that the use of these weapons reduces “blood, treasure and reputation” costs, the incentives to prolong conflict until the desired outcome is achieved could increase. Usually, as time passes and more dead soldiers are sent home, public opposition grows. Such was the case in the US during the Vietnam War and, more recently, as a result of the never-ending interventions in Afghanistan and Iraq. The financial costs of prolonged wars are, of course, another burden on the policymaker and another reason for public opposition. If LAWS were to tear down these obstacles, will we enter a state of “permanent war”?
However, what I have explained here is all theory. What is the real picture in practice? If you want to learn more about LAWS and how they can affect our international security from a practical and individual perspective, keep an eye out for my next posts!
Special thanks to Renée Gerboles, who wrote her final bachelor’s thesis on “The practical implications of the use of Unmanned Aerial Vehicles in warfare by Western States”.