We thank you for circulating the extremely useful summary of views expressed by various delegations on the characterisation of LAWS. We also thank you for your opening remarks outlining four possible approaches for the characterisation issue. It has greatly helped in structuring the discussions and channelling our thoughts.
The discourse on characterisation should not become an avenue for complicating the debate, nor for procrastinating while the relevant technologies continue to be weaponised.
My delegation sees the first approach, termed by you as the “separative approach” as the most appropriate to develop a characterisation of LAWS. By combining a process of exclusion and inclusion, under the “via negativa” and the “via positiva”, we can identify the key attributes that would characterise a given weapon system as LAWS.
This should be a technical and policy-neutral process. However, from what we have observed during the debate so far, it is being turned into a political issue suiting the respective policy positions, despite the caveat mentioned in your paper that: the issue of characterisation should be viewed independently and without prejudice to the eventual regulatory response. We should, first and foremost, de-link the characterisation debate from its policy and political dimensions.
For my delegation, the most important factor in characterising a weapon system as LAWS is the level of its autonomy. Weapons that can autonomously select and engage a target, also known as the weapons’ critical functions, without the direct involvement, control or supervision of a human, should be characterised as LAWS.
The idea is to capture all weapons systems that have the capability to act autonomously, regardless of whether their users have the intention to employ them in autonomous mode. Intentions should not guide the characterisation of and regulatory response to LAWS, but rather their inherent capabilities in terms of autonomy should guide this work.
We agree with Switzerland that autonomous weapons systems with manual over-ride or self destruct and self-deactivate functions should not escape their characterisation as LAWS. This also means that the concept of “meaningful human control” over autonomous weapons in of itself can neither provide the panacea to the challenges posed by them, nor lead to their exclusion from the scope of the GGE. This is again a capability vs. intention issue. A weapon system capable of acting autonomously regardless of whether the possessor intends to use it under human supervision should not affect our consideration of how we characterise it – independent of the eventual policy prescription that we might devise in the future.
At this stage we need to demystify and simplify the debate. We should try to reach on a simple common working understanding of LAWS without necessarily agreeing on a comprehensive and all encompassing legal-technical definition.
I thank you, Mr. Chairman.