[New publication] Balancing military and humanitarian interests: Scaling the scope of autonomous weapon attacks

In a new publication, researcher Jonathan Kwik proposes a scaling methodology to help characterise attacks by autonomous weapon systems (AWS). This could provide greater clarity on the legality of such attacks under international law, benefiting both civilians and belligerents.  

The new publication ‘The scope of an autonomous attack’ by Asser Institute researcher Jonathan Kwik addresses a crucial but often overlooked aspect of the debate on autonomous weapon systems (AWS): How broadly should we define ‘an attack’?  

Consider this example: At 12.00 hours, Commander-A activated an autonomous weapon system to attack a tank platoon, resulting in 7 shots fired at 4 tanks. Later, at 15.00 hours, the same system was deployed to attack 4 tanks guarding the city’s main routes, releasing 6 shots.   

How many attacks did Commander-A launch today?  

Scaling potential scopes of attacks 

International law regulates warfare by defining what constitutes an ‘attack’. However, there is currently no consensus on what exactly constitutes an autonomous weapon systems attack, leading to legal ambiguities and practical challenges.   

To address this uncertainty, Kwik suggests using a scaling methodology to determine an attack’s ‘scope’. This method helps military commanders decide the extent of autonomous activity that can be considered a single attack under international law, considering the proximity in time and space. Clarity on the 'attack scope' benefits both civilians and belligerents by balancing humanitarian and military interests.   

Kwik’s methodology shows that the scope of autonomous weapon systems attacks can vary widely. An attack could be defined as a single shot, a series of shots at the same target, bursts of shots targeting small groups, or a broader campaign against multiple objectives with a shared 'signature.' 

Attack scope spectrum 

According to Kwik, these varying definitions can be summarised into an 'attack scope spectrum,' including: 

• Shot: Each individual shot
• Single: All shots at the same target
• Group: Bursts of shots at small groups
• Signature: All shots against targets with a shared characteristic 

For example, if we define an attack by each shot, Commander-A launched 13 attacks. If we define it by tallying the number of targets, there were 8 attacks.  

Determining the appropriate scope for autonomous weapon systems attacks 

Before each attack, military commanders must conduct a legal analysis based on proportionality, precautions, etc. The chosen scope affects the level of involvement, creating trade-offs between protection and practicality. Humanitarian interests favour narrow interpretations, while military efficiency prefers broader ones.   

A broad interpretation (e.g., 'each activation') allows extensive autonomous activity, while a narrow one (e.g., 'each shot') could limit the use of autonomous weapon systems entirely. Kwik advocates for a middle-ground approach, balancing humanitarian and military needs. The appropriate scope for a single attack is determined by the proximity in time and space, which are context specific.  

Therefore, if we apply Kwik’s methodology to the provided example, we conclude that Commander-A has launched 5 attacks.  

Wondering how we reached 5 even though this does not align with any of the mentioned scopes? Find your answer by reading the full article and explore how the methodology can be applied in practice. 

About Jonathan Kwik 

Dr. Jonathan Kwik is a researcher at the Asser Institute under the ELSA Lab Defence project, who specialises in the laws governing the conduct of hostilities and artificial intelligence (AI). He recently obtained his doctorate (cum laude) from the University of Amsterdam on the lawful use of AI-embedded weapons at the operational level with a groundbreaking dissertation and is making waves in the legal and technical communities.  

Jonathan is part of the Asser Institute research strand ‘Regulation in the public interest: Disruptive technologies in peace and security’, which addresses regulation to safeguard and promote public interests. It focuses on the development of the international regulatory framework for the military applications of disruptive technologies and the arms race in conventional and non-conventional weapons. The public interest in peace and security serves as the prime conceptual framework in this strand.

Dr Jonathan Kwik