What is the value of sustaining international peace in the age of new technologies with incredible disruptive potential, such as the combination of weapons and artificial intelligence (AI)?
The financial costs of multilateral forums where these important issues are deliberated (such as the Group of Governmental Experts on lethal autonomous weapons systems I chair this year in Geneva) is nothing compared to the funds spent on national defence worldwide, or the potential funds that may be spent globally by a push to speedily modernise the world’s militaries.
Such a potential arms race would not only take away funds better spent on other social needs in many countries, but may impulsively hurry actors to give autonomy to weapons systems and subsystems prematurely, creating a recipe for instability. Sustained engagement with partners and adversaries alike — through multilateral deliberations — can slow such an impulse.
The value of multilateral discussions should more often be communicated for its real worth: it is the alternative to potential conflict. Multilateral diplomacy does produce international norms, but the processes of norm development and their maintenance — which is basically sustained communication — is equally important. Finding consensus around complex issues requires engagement, creativity and compromise.
Conflict can easily erupt due to misinterpreted intent. This is one aspect, among many, at which multilateral forums on security issues excel: they provide the opportunity to convey one’s own long-term intent and understand the intent of others. This certainly happens without these forums as well, but the real value in terms of avoiding conflict is achieved when there is a focused international deliberation, with the views of all stakeholders presented and taken into account.
The US National Security Commission on Artificial Intelligence released a consequential report last week. The document is impressive in its depth of analysis, but there is one aspect that makes it even more impressive: it is not a confidential document. Most major military powers have released some form of a strategy on autonomous weapons. Such documents are useful to deconstruct the incredible complexity that will undoubtedly affect future international relations and are critical inputs to the multilateral deliberations.
The report notes a 'strategic competition' centred around AI, but also 'deep interdependencies of the world’s leading AI states'. Among its conclusions is that 'the issues are too complex and vast for any part of [US] government, society, or industry to address alone'. The complexity, thus, of addressing the combination of AI and weapons internationally is multiplied by the number of states involved. This is why sustained international communication is essential, with views of all — bigger military powers, but also middle powers, smaller countries, and civil society — taken into account.
Machine learning, big data, and stronger computing power are enabling the adoption of ever-greater automated and autonomous functioning in existing weapons — adding degrees in speed, precision, efficiency, endurance, and power. Yet, as attractive as that potential may be, militaries are very serious about maintaining command and control over their weapons. Unless, that is, they worry that others may not be as careful. Such concern can fuel an arms race, as the drive for weapons modernisation is often driven by the (mis)perception of intent of potential adversaries. Given that speed is a critical factor, countries may worry that adversaries may create more reactive weapons and thus gain an upper hand, which may drive them to cede autonomy to weapons systems of sub-systems earlier than may be prudent, thus increasing potential instability.
This might make it seem as if it is impossible for there to be any international normative deliberation of any kind. Yet, the international debate in Geneva on autonomy in weapons has been constructive, focused, substantive, precise, and civil. The group has produced eleven Guiding Principles and great granularity of understanding, delving into the full complexity and security implications of autonomous weapons.
These deliberations are in danger of coming to a halt because of hampered financial liquidity of the Convention on Certain Conventional Weapons (CCW), under which they operate. The annual cost of the CCW — of which deliberations on autonomous weapons is but one small part — is approximately the cost of one Tomahawk missile, or about USD$1.5 million. The annual cost of the Biological Weapons Convention (BWC) is the same. This is to fund some of the most consequential deliberations in the world, and its annual budget is that of a small international NGO. Even with such a tiny budget, there are problems every year with financial liquidity.
Each of these conventions cost countries several thousand dollars per year; for many it is a few hundred; for a few countries it is in the tens of thousands and for a handful it is six figures. And both conventions are at risk of becoming dysfunctional due to hampered financial liquidity — the money is not there when it needs to be, since some countries do not pay at all, others don’t pay on time, and the conventions are not structured appropriately to handle the resulting financial gaps. Other UN bodies face similar financial challenges.
However, when I chaired the BWC last year, we solved this liquidity problem with a few simple tweaks. What was needed was for diplomats to think like accountants and for UN bodies with fiduciary responsibility to understand the concerns of national diplomacies. The same can be done for the CCW.
Very often, countries that internationally often hail the values of multilateral engagement do not make the effort to communicate its importance domestically, in internal political and administrative processes. The reason why countries are sometimes in arrears is simple bad communication between ministries of foreign affairs or defence and ministries of finance. Domestic communication, it seems, is as important to for maintaining multilateral forums for discussion and avoiding dangerous ambiguity that may result in conflict as is international communication between states. International and national communication builds trust through sustained engagement.
In our discussions on autonomous weapons, countries have shared their outlook, including their security concerns. National delegations consist of legal, technological and military experts who meet with their counterparts regularly. These deliberations have resulted in the most advanced international normative discussions compared to any other area in which the potential impact of new technologies based on AI is discussed.
New technologies based on AI are likened to a third historical revolution in warfare, after gunpowder and nuclear power. Only, current technologies are much easier and faster to replicate, which means they will hardly remain solely in the hands of major global powers that have the resources to develop them. This is one of the main reasons to continue having these multilateral deliberations.
Continued engagement and communication are better than working in secrecy; even during tensions, it is better to communicate in open forums than to cut off contact and assume the intent of others. Multilateralism is a financial commitment that, dollar-for-dollar, provides greater degrees of security than any other national budget line. As the world powers spend more on national security, and the UN bodies risk not paying their civil servants or holding meetings, we should remember the value that different arms control, disarmament, and non-proliferation regimes offer. The value is not simply the documents they produce, but also the opportunity to engage, negotiate, communicate, co-operate, and to send and receive direct and nuanced messages. This point must be a fundamental part of both national and international security deliberations.
The real costs of these conventions, and the crucial discussions they sustain, is minimal; however, the cost of isolation and the loss of forums for multilateral engagement may mean the demise of the international system as we know it.
What do you see as the dangers of using AI in warfare?
Join the discussion in the comments below!