Building an Ethical Framework for the Implementation of AI into the UN's Armed Conflicts Student Officer: Zoë Perkul Table of Contents I. Introduction II. Definition of Key Terms a. Artificial Intelligence (AI) b. Peacekeeping c. Lethal Autonomous Weapon Systems (LAWS) III. General Overview a. Development of AI b. UN Peacekeeping and Intervention IV. Major Parties and their Views a. UNICRI b. UNODA c. ICRAC d. United States of America e. Non- State Actors V. Relevant Documents a. Resolution 57/337 Prevention of Armed Conflict VI. Questions to Consider VII. Possible Solutions VIII. Conclusion IX. Bibliography I. Introduction Page 1 of 11
A world in which machines have the capability to form free, independent thought, is one that had previously only been seen in the distant settings of science fiction novels. Characteristics which had previously been solely associated with living, breathing creatures have now been adopted by machines. The ability to weigh options, make decisions, and speak freely, are all aspects of human life that are being made possible for automatons to perform. We use AI on a daily basis without the realization of such, applications such as Siri and Alexa are prime example of how we unwittingly rely on such automatons for simple tasks. Machines were created and are used as a means to make our lives more comfortable and effortless. We continue to use them under the assumption that they are not an immediate risk to us and are solely there for the betterment of those aforementioned lives. However, it is certain, as it has been in the past, that this technology and machinery will be used as a weapon in warfare or in deterrence. Whether it was gunpowder, gas, or nuclear weaponry, we have always weaponized it. However, the issue at hand is based on an ethical question: should the UN and member states be permitted to design and implement machines that are able to target and attack human beings without someone directing it to and controlling its every motion? Computer scientists have been coming increasingly close to developing such a weapon which has come to be called killer robots, stemming from AI s science- fiction roots. II. Definition of Key Terms a. Artificial Intelligence (AI) Artificial intelligence is defined as a branch of technology that has the capability of imitating human thought and intelligence. The Encyclopædia Britannica defines it as, the project of developing systems endowed with the intellectual processes characteristic of humans, such as the ability to reason, discover meaning, generalize, or learn from past experience. John McCarthy was the first to coin the term in 1956, when he held the first academic conference on the matter of AI. Artificial Intelligence has also been defined as the fourth industrial revolution. The first three other stages being steam power, electricity, and information technology. The fourth revolution is characterized as being a time driven by extreme automation and ubiquitous connectivity. This shows the magnitude of importance that artificial intelligence possesses for it to be seen as the fourth revolution, parallel to other technological advancements that have shaped the current world so immensely. Page 2 of 11
b. Peacekeeping The UN has a system in place for the active upkeep of international peace by deploying people to conflict- torn nations. The Peacekeeping program was set up with the intentions of providing security and the political and peacebuilding support to help countries make the difficult, early transition from conflict to peace. Alongside this, they also work facilitate the political process, protect civilians, assist in the disarmament, demobilization and reintegration of former combatants; support the organization of elections, protect and promote human rights and assist in restoring the rule of law. Currently they have operations that span 15 conflict ridden nations. c. Lethal Autonomous Weapon Systems (LAWS) Lethal Autonomous Weapons are military robots which have the capability to be tasked with selecting and attacking targets specific to the military, such as people or installations, without a human operator in charge of the machine. While these systems have not yet been implemented in warfare, many developers have been making significant strides in the creation and development of such machines for the purpose of combat. However, an onslaught of computer scientists has warned the UN of the potential risks that LAWS could pose to the security of nations. They claimed that through the continuation of the implementation of AI, they could potentially be opening Pandora s Box. Some have stated that LAWS could be the spark which could ignite the next arms race between nations, with the number of fatalities that it could cause rivaling any and all acts of war we have experienced in the past. Often times, in the media, they are being referred to as killer robots, a simplified definition of what their purposes may be. III. General Overview a. Development of AI The world of electronics has been changing and advancing at an extremely rapid pace since the invention of the computer in 1936. In 1970, Gordon E. Moore, a computer analyst, created an idea which would later become known as Moore s Law, which states that the processor speeds, or overall processing power for computers will double every two years. Although this theory has been heavily discussed and debated, it is still accepted amongst many technicians, however, it is coming to an end as the physical limit of the microprocessor is being reached without similar breakthroughs in Quantum Computing This notion portrays the exponential development that is made, generally in technology, in such as small number of years and how it will theoretically continue to do so infinitely. Page 3 of 11
AI was initially only something thought about and discussed, something far from a reality until the 1950s when it was first created under completely innocent motives. AI found its first large foothold within the realm of computer chess from which the notion of computers making semi- independent decisions as a reaction to something occurring became a reality. In more recent years, Google s AlphaGo program emerged in 2015, which is an AI program that was developed to be able to play the ancient Chinese game Go. This is a game that has been played for thousands of years and due to this there are people who have mastered the game and were thought to be unbeatable. However, the AI program was able to beat all of them and is now the reigning world champion in that game and has yet to lose. The program used an algorithm, however, experts claimed that it made moves that were so advanced and unimaginable which had never been seen before. The developers of AlphaGo had this to say about the programs accomplishments, We have always believed in the potential for AI to help society discover new knowledge and benefit from it, and AlphaGo has given us an early glimpse that this may indeed be possible. From that point on, researches became aware of the potential effects that artificial intelligence could have in many aspects of life, particularly in warfare. They realized that if they were to create completely autonomous lethal weapons then they would be able to remove any possibility of human error. Through this they would also be able to cut down on the amount of budget they put towards UAV pilots. However, with this realization, many brought up the ethical and moral questions posed by this advancement in technology. They argued that the development of AI could have many disastrous implications for human kind, such as; unemployment, inequality, effects on our humanity, and security. It is due to these objections that the ethical question is being debated on regional, national, and international scales. A new ethical question has been brought to light which deals with the ethics that AI possesses, or at least what it has been programmed to possess. With the creation of the self- driving car came the moral and ethical question of the Trolley Problem. The Trolley Problem goes as follows, There is a runaway trolley barreling down the railway tracks. Ahead, on the tracks, there are five people tied up and unable to move. The trolley is headed straight for them. You are standing some distance off in the train yard, next to a lever. If you pull this lever, the trolley will switch to a different set of tracks. However, you notice that there is one person on the side track. You have two options: do nothing, and the trolley kills the five people on the main track or pull the lever, diverting the trolley onto the side track where it will kill one person. Which is the most ethical choice? This problem has been raised in the context of when affronted by the situation, would the car choose to harm five pedestrians or the driver. It is thought that the programmers essentially choose for them as they encode it in their logarithms. This further highlights one of the ethical objections people have against AI, that they have, in some cases, the ability to choose who lives and who dies. Page 4 of 11
The field of artificial intelligence and computer sciences is an ever- changing concept, one that is especially current in our globalized world which has brought issues stemming from a national scale to an international level. The speed at which it all develops is one that is so exponentially fast that one must stay up- to- date on all of the advancements that are made on a daily basis. It is difficult to predict what changes will be made in this field, however, where it concerns warfare it is easy to say that any advancements made in the future will only build off of the concept that has already been stated; a machine that is built for the sole purpose of war, which has taken all aspect of humanity out of the equation. b. UN Peacekeeping and Intervention Chapter Seven of the UN Charter outlines what is to be done, within the UN s power, in cases of threats to international peace and unsanctioned acts of aggression. It states that in urgent and dire cases the UN has the ability to intervene through most any means they deem suitable, be that demonstrations, blockade[s], and other operations by air, sea, or land forces of Members of the United Nations. Due to the aforementioned ability of military intervention, one must consider the UN s relationship to AI. Although many desires for the prohibition of artificial intelligence in warfare have been voiced, whether or not this is going to be put into effect is solely a matter of waiting. However, it is also important to examine whether the implementation of AI into warfare, both on behalf of member nations and the UN s peacekeeping initiatives, would be impeding upon the UN s core mission of maintaining international peace and protecting innocent civilians in conflict torn nations. As previously defined, the UN s Peacekeeping missions are used to preserve and promote international peace by using armed conflicts prevention methods. They do so using a wide array of methods, including; the use of military, police and civilian volunteers. In particular, the use of the UN military, often referred to as Blue Helmets, is done so in cases of conflicts where the UN feels that the local citizens and or borders, for example, require monitoring and protection. The UN military has been used in cases where there are large conflicts taking place, such as the Middle East and central Africa. IV. Major Parties Involved and their Views a. UNICRI The United Nations Interregional Crime and Justice Research Institute (UNICRI) is an UN research facility that specializes in the spread of crime from national to an international spectrum. Within that, they also focus on the emergence of new forms of crime and weapons, Page 5 of 11
such as LAWS. They also handle issues such as; chemical, biological, radiological, and nuclear risks mitigation, cyber- crimes, and protection of vulnerable population and victims. In 2015 they launched an initiative called Chemical, biological, radiological and nuclear (CBRN) National Action Plans: Rising to the Challenges of International Security and the Emergence of Artificial Intelligence to foster co- operation and co- ordination both between different national agencies and among countries and international and regional organizations. This is due to the fact that managing criminal risks, both of accidental or natural origin is extremely difficult but also of the utmost importance. This international co- operation is used for the purpose of establish[ing] common goals and identify[ing] and manag[ing] resources to achieve them; clearly allocate[ing] responsibilities and tasks; elaborate[ing] functioning channels of communication; creat[ing] a security culture based on common learning; and ensur[ing] that lessons learned are incorporated and absorbed throughout the whole system. The crowning achievement of this initiative was the implementation of National Action Plans regarding the mitigation of criminal risks, including the development of AI as a weapon of warfare. These have been implemented in over fifty nations internationally. b. UNODA The United Nation Office for Disarmament Affairs has taken substantial steps towards the prohibition of fully autonomous weapons in warfare. In the year 2017 they collaborated with NGOs such as International Committee for Robot Arms Control, Human Rights Watch, Seguridad Humana en Latinoamérica y el Caribe and the Campaign to Stop Killer Robots in an effort to fully realize the potential effects AI weapons could have on international security. On the 16 th of October, 2017, the UNODA hosted a panel to the topic of Pathways to Banning Fully Autonomous Weapons in cooperation with the Permanent Mission to the United Nations of Mexico and a host of other NGOS as a means to foster fruitful discussions about the possible changes AI could bring to war and the repercussions they could have upon us as a society. In this discussion, the Ambassador for the Permanent Mission to the United Nations of Mexico stressed the importance of continued research and education on these weapons systems and their possible dangers. c. ICRAC The International Committee for Robot Arms Control is an active participant in discussions concerning the prohibition of LAWS. In their mission statement it reads that one of their main objectives as an organization is the prohibition of the development, deployment and use of armed autonomous unmanned systems; machines should not be allowed to make the decision to kill people. ICRAC had an extremely influential role in establishing the movement Campaign to Stop Killer Robots, alongside the Human Rights Watch, amongst Page 6 of 11
others, which is an initiative which has been fighting for the ban of all autonomous unmanned systems for the past five years. The campaign has consistently been arguing that Human control of any combat robot is essential to ensuring both humanitarian protection and effective legal control. d. United States of America The United States of America has had the largest and most advanced military for a number of years, while also being more advanced in most military fields than any other nation. In 2015, the Pentagon predicted that the annual military budget was $598 billion, of which $2.4 billion is being spent on research, upgrades, and procurement of unmanned aerial systems. The USA has, since 2001, been implementing unmanned aerial systems in combat due to the fact that they are cheaper and have a higher longevity than planes. However, the use of artificial intelligence in the unmanned aerial vehicles (UAV) applications has not yet been enacted. A multitude of AI companies, which have already created UAVs which use AI, claim that their AI now easily bests highly trained pilots in simulated aerial combat. Currently, the USA has plans to use AI drones as a means of surveillance on the battlefield, due to the fact that the moral and ethical questions surrounding AI are put aside for this purpose as it does not directly harm any human beings. e. Non- State Actors As of right now only multi- national companies have been able to manufacture AI and the associated technology, however, how long will that be able to continue for? One of the largest benefits of the digital revolution is that it is relatively cheap in comparisons to the past three revolutions, which could potentially allow non- state actors to participate in the developing and manufacturing of AI technology. During the arms race of the cold war, only nations with extremely rich military programs could develop nuclear weapons which was then used for the military tactic of brinkmanship. However, if non- state actors are able to also manufacture lethal AI programs, then the possible ensuing arms race would be unimaginable in its scale. The way the arms race ended in the 1970s would no longer be possible because it would not solely be governments involved but rather individuals and companies too. V. Relevant Documents a. Resolution 57/337 Prevention of Armed Conflict This resolution, which was passed and adopted by the General Assembly on the 3 rd of July 2003, outlines the responsibilities that the member states have agreed to regarding the prevention of Page 7 of 11
any potential armed conflicts. This document is of the utmost importance due to the fact that it recognizes all the factors that could possibly lead to such a conflict and emphasizes the need for international dialogue. Encourages greater transparency in armaments by Member States [ ] relating to arms registers and military expenditures, and calls upon them strongly to support confidence- building measures in this area. Alongside this, it calls for the implementation of all measure previously outlines by all treaties and resolutions regarding arms control and non- proliferation. Also called for in the resolution is the strengthening of freedom, justice, democracy, tolerance, solidarity, cooperation, pluralism, cultural diversity, dialogue and understanding as these are important elements that contribute to maintaining peace and deterring the threat of conflict. VI. Questions to Consider There are some aspects that all delegates must consider before moving forwards in their discussions and debates and those are as follows; Could the implementation of AI into any member states military infringe upon Article Three of the Universal Declaration of Human Rights? How could we prohibit specific advancements in certain fields of AI technology while still encouraging advancements in other technological fields? How would a total ban of fully autonomous weapons be implemented in such a way that can be monitored and verified? How could AI UAVs pose a threat to international security? Due to the fact that a complete prohibition is not the only possible solution, could a legal framework be put in place to regulate LAWS? If so, what should be detailed within the framework? Could AI also pose potential risks such as cyber- terrorism? What effect could it have if non- state actors are able to use AI to create weaponry? VII. Possible Solutions Due to the fact that this is an ongoing and very contemporary issue, all solutions must take into account the advancements that are being continuously made in the field of AI technology. The call for a complete ban of LAWS has been increasingly requested by both laymen and computer scientists due to the catastrophic effects they could have upon the human population. While the UN has held conferences and has entertained this as a disarmament committee topic, no definitive steps have yet been taken. However, it is Page 8 of 11
important to consider that any limitations put upon the development of artificial intelligence must be carefully thought out due to the potential benefits AI could have upon the society. Banning AI in its entirety could severely limit the possible unforeseeable future advancements that could be made in a host of fields that greatly benefit the international society as a whole. One must also take into account that a ban put upon any AI developments may not be entirely effective as one would hope. The idea of introducing a ban on anything is very difficult to fully enforce and monitor and seeing as it is not only member states that are acting in the development of AI weaponry, but also non- state actors, this would only work towards increasing the difficulty. Delegates must take into consideration that there are also alternatives to regulating the use of fully autonomous weapons without implementing a complete ban on them. By creating a legal framework to outline the exact limitations put upon certain aspects of LAWS advancements, it may be a more realistic and possible approach to the issue. It would then be at the delegates discretion to decide what exactly these limitations would be and how they plan on enforcing such decisions. VIII. Conclusion The current issue of building and regulating an ethical framework for the implementation of artificial intelligence into the UN s conflict prevention agenda and, as an extension, that of all member states, is of a complexity that is difficultly managed. Technological innovations have propelled many fields into the next step of their evolution. An example of this is medicine, technological advancements have made formerly impossible tasks possible and have made previously incurable diseases curable. However, the next step in AI in warfare may not be the solution to costs and human error as some war theorists think it may be. Important aspects have to be taken into consideration. These include the realization that automatons could eventually have the ability to choose who lives and who does not on the battlefield, thus being in complete control over their life. While negative advancements in AI are easy to point out, there are also a host of positive impacts it could have on the society. With this in mind, one must consider if one of these sides outweighs the other. IX. Bibliography Copeland, B.J. Artificial Intelligence. Encyclopædia Britannica, Encyclopædia Britannica, Inc., 23 Feb. 2018, www.britannica.com/technology/artificial- intelligence. Page 9 of 11
A New Dawn. Artificial- Intelligence, UBS, www.ubs.com/microsites/artificial- intelligence/en/new- dawn.html. Smith, Chris, et al. The History of Artificial Intelligence. University of Washington, Dec. 2006, courses.cs.washington.edu/courses/csep590/06au/projects/history- ai.pdf. Killer Robots: Experts Warn of 'Third Revolution in Warfare'. BBC News, BBC, 21 Aug. 2017, www.bbc.com/news/technology- 40995835. Pathways to Banning Fully Autonomous Weapons. United Nations, United Nations, 23 Oct. 2017, www.un.org/disarmament/update/pathways- to- banning- fully- autonomous- weapons/. McSweeney, Kelly. UN to Host Discussion about AI in Warfare. ZDNet, ZDNet, 13 Nov. 2017, www.zdnet.com/article/un- to- host- discussion- about- ai- in- warfare/. What Is Peacekeeping. United Nations Peacekeeping, United Nations, peacekeeping.un.org/en/what- is- peacekeeping. United Nations General Assembly. 57/337. Prevention of Armed Conflict. UN Documents, United Nations, www.un- documents.net/a57r337.htm. Chapter VII. United Nations, United Nations, www.un.org/en/sections/un- charter/chapter- vii/index.html. CBRN National Action Plans: Rising to the Challenges of International Security and the Emergence of Artificial Intelligence. UNICRI, United Nations, www.unicri.it/news/article/cbrn_artificial_intelligence. Moore s Law. Moore's Law, www.mooreslaw.org/. Walker, Jon. Unmanned Aerial Vehicles (UAVs) - Comparing the USA, Israel, and China. TechEmergence, 1 Sept. 2017, www.techemergence.com/unmanned- aerial- vehicles- uavs/. 67th Session of the Secretary- General's Advisory Board on Disarmament Matters Discusses the Impact of Artificial Intelligence on International Security. United Nations, United Nations, Page 10 of 11
www.un.org/disarmament/update/67th- session- of- the- secretary- generals- advisory- board- on- disarmament- matters- discusses- the- impact- of- artificial- intelligence- on- international- security/. Bossmann, Julia. Top 9 Ethical Issues in Artificial Intelligence. World Economic Forum, World Economic Forum, www.weforum.org/agenda/2016/10/top- 10- ethical- issues- in- artificial- intelligence/. AlphaGo. DeepMind, DeepMind Technologies Limited, deepmind.com/research/alphago/. Artificial Intelligence and Robotics. UNICRI, United Nations, www.unicri.it/topics/ai_robotics/. Military Peacekeeping. United Nations, United Nations, peacekeeping.un.org/en/military. About ICRAC. ICRAC, 22 Feb. 2018, www.icrac.net/about- icrac/. Marshall, Aarian. Lawyers, Not Ethicists, Will Solve the Robocar 'Trolley Problem'. Wired, Conde Nast, 1 Feb. 2018, www.wired.com/2017/05/autonomous- vehicles- trolley- problem/. The Solution. Campaign to Stop Killer Robots, www.stopkillerrobots.org/the- solution/. About UNICRI. UNICRI, United Nations, www.unicri.it/institute/. Page 11 of 11