Autonomous Weapons and AI-Enabled Military Systems

General AssemblyDisarmament & International Security Committee

TopicAutonomous Weapons and AI-Enabled Military Systems

To many, the term “autonomous weapon” often evokes images of flying drones or unmanned aerial vehicles (UAVs). While there is truth to this impression, the term is actually much broader. There is no official international definition of autonomous weapons, though many international organizations classify them as weapons that apply force without human intervention. Traditional drones require a human operator, someone whose remote input guides their every action, and who has the final say on whether or not to apply force. Autonomous weapons either limit or entirely lack this human oversight, especially when it comes to the application of force. One of the facets of the problem is that there is a sliding scale of autonomy. If an advanced weapons system is empowered to provide a list of potential targets of attack, there is a degree of autonomy that some nations may find troubling, even if the final decision to use force still rests in human hands. Most analysts of the problem agree that there should be what is called “meaningful human control,” but various stakeholders differ on what exactly that means.

The other concept DiSec is grappling with in this topic is so-called “artificial intelligence,” or AI. The term itself is something of an exaggeration. True AI that is self-actualized, has its own thoughts and personality, and could be said to have a real claim to consciousness does not yet exist. The term AI in its present usage typically applies to various implementations of machine learning, which involves passing a great deal of data through complex computer algorithms in order for it to ultimately generate responses or “make decisions.” These can take the form of large language models (LLMs) such as ChatGPT, Claude, or Gemini, used by individual consumers for private or business use, or more specialized machine learning tools for government or military use. While AI and its capabilities have drastically improved in recent years, the conclusions drawn and decisions “made” are ultimately still the product of a computer with something less than genuine consciousness, and in any case at a remove from human judgment.

The regulation of AI use in autonomous weapons and military systems has rapidly become a global priority. UN Secretary-General António Guterres has called for a global ban on Lethal Autonomous Weapons Systems (LAWS), though no such measure has yet been implemented. The High Contracting Parties to the Convention on Certain Conventional Weapons (CCW), a treaty responsible for restricting and banning weapons that cause unjustifiable or unnecessary harm to civilians or combatants, have called for a Group of Governmental Experts (GGE) to meet and discuss potential restrictions for autonomous weapons systems. These discussions have produced ideas without implementation. In order to successfully address these issues, it’s important to consider member states’ capacity for the development and use of autonomous weapons and/or AI. The creation and use of such systems are resource-intensive processes, meaning not every nation will possess the capabilities to maintain programs of their own. It will be important to consider not only how countries are allowed to develop and use these tools, but how the information pertaining to them can be shared between countries.

Focus Questions:

1. Should the usage of autonomous weapons be restricted to certain areas or purposes, or banned entirely?
2. Is there an acceptable level of autonomy in advanced weapons systems? How should that level be defined? What constitutes “meaningful human control?”
3. Who should be responsible for regulating the usage and development of AI? Individual governments, the UN, another intergovernmental organization (IGO)?
4. What restrictions, if any, should exist on trade in AI and autonomous weapons, or the intellectual property related to their development and use?

Useful Links:

UN Institute for Disarmament Research (UNIDIR) Working Papers on LAWS and Military AI:
https://unidir.org/publication/unidir-on-lethal-autonomous-weapons/
https://unidir.org/publication/artificial-intelligence-beyond-weapons-application-and-impact-of-ai-in-the-military-domain/

Presentation to the Informal Meeting of Experts on LAWS of the CCW, concerning “Meaningful Human Control”:
https://docs-library.unoda.org/Convention_on_Certain_Conventional_Weapons_-_Informal_Meeting_of_Experts_(2015)/BREHM_Presentation%2Bon%2BMHC_14.04.2015.pdf

GA Resolution on LAWS:
https://docs.un.org/en/a/res/78/241

Submit a Position Paper

Submitted Position Papers