2026 Group of Governmental Experts on Lethal Autonomous Weapon Systems
First Session (Geneva, 2–6 March 2026)
Asia-Pacific Institute for Law and Security
Utrecht University
Thank you, Chair!
I am delivering this intervention on behalf of the delegation of the Asia-Pacific Institute for Law and Security (APILS) and my own delegation, Utrecht University.
First, we would like to thank you and your team for your tireless work in getting the text to where it is today. We also thank you for your engagement with civil society and academia during the intersessional period.
Chair,
We support the thrust of Box I on the characterisation of LAWS.
Almost every word in the characterisation reflects several years of painstaking discussions. It captures the meaning of LAWS with considerable precision yet has the virtue of remaining concise. Subparagraphs A and B helpfully clarify some elements of the characterization that several delegations have found critical. Those subparagraphs wisely abstain from abstract definitions of notions such as “identify” and “lethality” but focus on the scope of the characterisation.
Some delegations raised questions about the phrase “functionally integrated”, which we have previously supported. The challenge here lies in ensuring that the characterisation of LAWS captures “weapon systems” rather than simply “weapons”. The word “functionally” simply ensures that the “combination of one or more weapons and technological components” does not need to be integrated physically into a single device, but must be integrated into a single weapon system in terms of the functions of individual technological components. It is this integration of functionality that allows the weapon system to “identify, select, and engage a target, without intervention by a human operator”.
With respect to subparagraph A, we support the suggestion by the ICRC’s to replace the reference to “the parameters of the target(s)” with a broader reference to “the list of potential targets, the generalized target profile, or the operational parameters of the system”.
An alternative and more comprehensive formulation we would propose for this paragraph would be: “The fact that a human operator sets the parameters that comprise the profile of the target(s) does not exclude the system from the characterization”.
Chair,
We continue to have concerns about the phrase “within the scope of application of the CCW” included in the square brackets. For one, it is unclear why it is necessary, considering that the GGE operates under the auspices of the CCW and within the mandate given to it by the Meeting of High Contracting Parties.
But perhaps more importantly, the expression “scope of application” has a specific technical meaning under Article 1 of the CCW. This Article defines the “scope of application” of the Convention as the circumstances in which the Convention applies as a matter of law – namely situations of international armed conflict and, following the amendment to Article 1, non-international armed conflict. Using this phrase in the present context therefore imports a treaty concept that is tied to the issue of the application of the Convention.
Put differently, the characterisation of LAWS risks becoming enmeshed with the question of legal characterisation of armed conflicts. This is not how the CCW and its Protocols have traditionally regulated weapons. The Convention regulates weapons based on their characteristics — that is their potential to cause superfluous injury or unnecessary suffering, or have indiscriminate effects. The current wording, however, appears to unnecessarily link the characterisation of the weapon to the circumstances of its use rather than its inherent features.
We suggest therefore that the phrase “within the scope of the application of the CCW” be deleted.
Furthermore, the phrase “for the purposes of these elements of an instrument” is perfectly sufficient to limit the application of the characterisation to the present document. This more concise language would also mirror the simple phrase “for the purpose of this protocol” used to introduce definitions in Article 2 of Protocol II and Article 2 of Protocol V.
We also consider that the phrase “can be characterised as” creates some ambiguity. The word “can” suggests that an AWS could be characterised in the way contemplated in the rolling text, but that it could also be characterised in some other way — possibly in a manner that has not been considered by the GGE. We submit that the simple word “is” would be sufficient to replace the phrase “can be characterised as”.
Thus, the beginning of the paragraph could read: “For the purpose of these elements of an instrument, a lethal autonomous weapon system is a functionally integrated combination …” and so on. This would not prevent, of course, the GGE from amending this characterisation later. I will submit our comments directly to your team after this intervention.
I thank you, Chair!
