7th Meeting - 1st Session Group of Governmental Experts on Lethal Autonomous Weapons Systems 2025

Excellencies, distinguished delegates, let me thank you for your understanding that we start a little bit later. It allowed me to do some further internal consultations and coordination with regard to the text we have discussed. over the last few days. As promised at the end of our meeting yesterday, we circulated some text proposals and we will put them up on the screen. Is that possible now? Thank you for that. Good. Let me say that I think that we can look back at three days of intense and fruitful discussions during which we completed a full reading of the rolling text. And as I expressed yesterday, I would also like to underline my sincere gratitude for the commitment, flexibility and constructive approach demonstrated by delegations. However, I think we have to go the extra mile. After all, we still have two full days ahead of us. And as announced at the closing of yesterday's session, you have received through the Secretariat a revised version of boxes one and two. These documents, let me stress that, have no formal status and will not be put on the website as they are issued under my sole authority as Chair with the purpose of facilitating our discussions today. You will have noticed some suggested changes in the boxes, which are my best guesses at ways in which we might be able to bridge divergences across the room with regard to the rolling text. Let me first take you through these suggested changes that I propose in box one, which would now read as follows. Within the scope of the application of the CCW, A lethal autonomous weapons system can be characterized as an integrated combination of one or more weapons and technological components that can identify and/or select and engage a target without intervention by a human user in the execution of these tasks. Let me explain what we have done. We have added within the scope of the application of the CCW. This is to reaffirm that this characterization falls within the scope of the application of the CCW, a matter that several delegations mentioned that is important to make explicit of the draft set of elements. With regard to integrated combination, there was quite some discussion on this. While I think we already agree on the substance, it's a functional integration, not necessarily a physical one. In case this remains unclear, I invite delegations to suggest text to resolve this matter. We changed enable the system into can. This is a simplification streamlining to make clear that not only the technological components, but the system as a whole is capable of identify and/or selecting and engaging a target. Speaking about identify and/or select and engage, I have deliberately refrained from making suggestions on this. Because I first would like to continue the discussion with delegations on this paragraph. In particular, I would like to invite delegations to reflect on identify and elaborate further on their understanding of identify specifically in relation to the other functions. And on the question whether we really need this as a threshold in a broad working characterization. Because having identify, together with select and engage, especially in combination with two times and, the so-called cumulative approach, would significantly narrow down the scope of the characterization. Is this really necessary since we also have para 3 explicitly outlining that a description is without prejudice to any future understanding and potential modification as well as the possible exclusion of certain types of systems? It might help to get a view from delegations with the aim to get a better understanding of the autonomous process. Also, by having a short discussion on the kind of systems that delegations would want to include or exclude by using or avoiding this term, and it could possibly help to reach a common understanding. Then, with regard to in the execution of these tasks, I decided to retain this for now, since several delegations have requested to keep a reference to the temporal dimension of autonomy. The next para, para 2 now, is a new para and aims to resolve the issue of lethality. During our discussion, There were delegations who indicated they would like to retain lethality as a qualifier of autonomous weapons systems, because it's also in the mandate, and there were those who would like to see it removed from the characterisation. The compromise might be to retain it with a clear explanation. In this context, I would also again like to remind delegations that we are working on a broad working characterisation and not a narrow prohibition. Having said that, I would now like you to consider the following text based on a combination of suggestions received from delegation, and I would read them as follows. Para 2, for purposes of these elements of an instrument, lethal means weapons with the capacity to kill or injure one or more persons which includes weapons that can be used to damage or destroy objects. The third paragraph remains unchanged for now. I feel that the current wording of the exclusion clause gives ample room to modify the characterization in the future. So for the time being, it would continue to read as follows. The above description is without prejudice to any future understanding and the potential modification of this characterization, as well as the possible exclusion of certain types of systems. This concludes my introduction to box one. With this, I would now like to open the floor to hear from delegates before we move on to box two. And I would also like to encourage you to try to engage with each other and address or suggest possible compromise solutions. Depending on the flow of the discussions, we might also suspend the meeting to give delegations time to work in smaller groups on the different proposals and su- and suggestions that are on the table. Let me now open the floor. I recognize the distinguished representative of Austria, and the Chair is most grateful for him being the one to kick off this debate. Let me again assure you, if you have an early intervention, you always have the right to come back later on to react to the others, so you are not giving away any advantage to colleagues. In fact, you might shape the discussion. But now, let me pass the floor to the distinguished representative of Austria. You have the floor.

And this is mostly to break the ice because I think we are happy with many of the changes that you have made, except for some detail. In general, maybe also to support some of the comments that you have made, and we have said this before, we have to keep in mind that this is a working characterization, and every working characterization has to be as broad as possible, excluding at this point in time several systems just because there is a certain fear that there might be additional workload or problems in the later part of the text, or that they might fall under prohibition where states don't want to see them. I think this is the wrong approach, especially at this point in time. A broad characterization is the one that does not exclude anything. And in this context, you also have to remark that we are now in the 10th year of the treaty laws or in the 10th year of the discussion in the CCW, to be more correct. And if we now try to have the scope of our discussion, this would be a very negative step. And there is a lot that needs to be included into our work and the separation of certain elements or certain… ideas about excluding existing weaponry because they fully comply already with IHL. This should happen at the very end, in particular because there are many existing weapon systems that, with the text as we are seeing it right now, would not pose any problems, would fully comply with the requirements that we are setting out. We should also not forget, I mean, this is important, that many of the must requirements or other requirements that we are out in the different boxes have a basis under existing IHL. So there is no, not much reason for fear, and especially when it comes to the existing weapon systems, and I just wanted to make that point again. On the first para, And within the scope, I mean, we have seen this previously as a chapeau, but we have no problem including it here, especially if it makes it more specific. In our view, making clear that we are here working within the scope of application of the CCW would even be a good argument to exclude the lethal, because then it's clear what we are talking about. But we are seeing the other compromise that you have put on the table, but just to make this comment. the integrated combination of we hear your explanation and your interpretation of this wording would actually resolve probably the problem that we are seeing. But as usually there is the question, is everyone in this room interpreting it the same way as you are interpreting? So maybe we can have this discussion and that will make it easier for us to work with this wording. But this is still something that we need to hear from other delegation. On the identify and or, uh, select, um, we could live with the current form- formulation because it, uh, puts the focus on the select, um, on the identify, we are having the same questions that you, uh, put to the floor, um, we are not, do not fully understand what this exclusion… And we have to be clear, this what is called now the cumulative approach that was, I think, originally it was intended to be a negative term, and now it's embraced by many. I think what it does, it's limiting the scope of what we are discussing here. And the question is why? We know that many delegations would prefer to exclude systems that are already existing or that maybe are under development. But we are not completely clear which systems are we talking about, what kind of functions do they have that would not render them being an autonomous weapon system, and for what kind of reason is this the case? We are having some trouble understanding, and maybe it would be good to hear some concrete examples of what what states want to exclude if they identify. If we look at it, I mean, the nearest possibility is that they want to exclude anything that does not have a target identification system. But in our view, this would be highly illogical. There is a range of weaponry that does not have this kind of sophisticated capability, but by logic, by technical capability, cannot be defined as anything else as an autonomous weapon system. So this is a discussion also still to have. Um, I think this is a good combination about from the ICRC proposal with also some, some other elements. Um, the only point that we are seeing is that maybe it's not the best way to define lethal, um, through the way of a weapon, but to define lethal as a capacity. And the easiest fix would be to just delete weapons with the, so that it reads lethal means the capacity, uh, to kill or injure, so that we are not defining a capacity through the, uh, weapon. detour of a weapon, but this is a more grammatical issue. Once again, this is a working method that we have now found here in the room. It makes an explanation what we are doing here. The question is if we even have to say for the purpose of an instrument or we could even say for the moment for the purpose of the work within the GGE, but we are open to that. But for us, this does not necessarily mean that this is also a formulation that needs to go into the final instrument because this is also a general approach to the characterization box. It very much and strongly relates to the rest of the text, and we might come back to it at the very end. And I think this is also the approach other delegations are having when they want to look into exclusions or other elements that are mentioned in para 3. And this is all from our side. Thank you.

Let me thank the distinguished representative of Austria for sharing his views and making comments on the text proposed. Next on my list is Brazil, to be followed by Cuba. But let me now pass the floor to the distinguished representative of Brazil. Sir, you have the floor.

Thank you, Chair. And just for the sake of engaging in this interesting debate. I would also like to thank you for this alternative formulation. I think that the approach you followed is a sound one, which is to avoid changes, too drastic changes, and only advance when there is a clear possibility of narrowing down the areas of disagreement. I do think that the changes are positive. especially with regard to the clarification in the beginning of paragraph one. And on the concept of lethality, I think that this formulation assuages a lot of the concerns that we heard in the initial rounds of discussions. Of course, there is still some room for minor improvements and tweaks, but I think that this addresses the lack of clarity of the term lethal that we had going into this discussion, and I think that now we have a lot more clear vision about what we're talking about. I'd also like to support the comments made by my colleague from Austria regarding the bits that you clearly outlined as meriting further consideration, which is the reference to critical functions in paragraph one. My delegation is convinced that the best way to frame those critical functions is to limit ourselves to select and engage for the same reasons that Austria has outlined. And I think that the key to solving this problem is to have the delegations that is to hear from the delegations that are attached to the retention of identify a clear explanation about what they understand by this term, what is identify. And not only a clear explanation of the term, but also concrete examples of what they feel could be the consequences of including or excluding this as a critical function. I think the time is past for just general expressions of preference. We are not here to deal with national preferences. We all have ours. We need to understand clearly what is behind our discussions and the concrete terms, what the consequences of our choices will be. So very clearly, I think that select and engage is the way forward. Many of the of the elements for that underpin this assessment are very much out in the open and I hope that we will hear more in during our discussions at this at this stage. And again, as Austria has outlined, there is this is not a definition of what is to be prohibited. It's quite clear that it's not the case. Just a cursory reading of sections three and four would dispel any such notion. We are not defining what we will prohibit. If so, even my delegation would have a problem with it, but that's not the case. So please read this. paragraph in conjunction with what we are doing in the further boxes. That said, Mr. Chair, thank you very much to you and your team, and we look forward to engaging further in these discussions. Thank you.

Let me thank the distinguished representative of Brazil for commenting on the proposals made and also reacting to some of the earlier intervention of the Austrian delegation. Next on my list is Cuba, to be followed by Ukraine, Pakistan and the Asia Pacific Institute. But let me first pass the floor to Cuba. Madam, you have the floor.

Thank you, Chairperson, and we thank the Chair and friends of the Chair for the work done to produce these two boxes. Overall, we can align with the track that has taken in the first draft. However, we would suggest some small amendments and In particular on 2, the characterization, lethal should refer not to weapon but to capacity, because other weapons can have the capacity to be lethal. And in 3, we also believe that as well as the possible exclusion, that clause should be deleted. We are reviewing the draft and we will be sending a suggested text. This deletion would not change the definition and gives scope for a broader interpretation on one This is the most complex paragraph. In addition to critical functions, for us that is crucial, and this could include how the... We are concerned about whether the... about the issue of the human user in the execution of the tasks. There needs to be a human within the life cycle in order to bring responsibility under international law, both for human beings and states. We should be clear that, despite the intervention of the human user, we should be saying that, despite human intervention, there is responsibility under IHL, because otherwise we have no scope for that and there is ambiguity produced. So those would be our comments at this stage as regards this text. Thank you.

Let me thank the distinguished representative from Cuba for sharing the views of her delegation and making proposals. Next on my list is Ukraine, to be followed by Pakistan. But let me first pass the floor to the distinguished representative of Ukraine. So you have the floor.

Thank you, Mr. Chairman, and thank you for your effort in reaching to compromise. Ukraine delegation in general agree with this new version of this chapter. Uh, some, uh, comment regarding, um, terms select and identify. I sent to secretariat proposals in, uh, e-mail regarding replace, um, position of this, uh, category because from military point of view, short operation should be select. This term understand that from different object on combat field, some should be take as target. Particular with using artificial intelligence for this task can be used object detection technology on board of LAFS. In this case, it will be build bounding boxing for different targets. After this, the system should be identified. It is enemy or blue forces object, as example, IFF function. It is civilian or military person. This target was destroyed before or not and should be defeated. It should be defeated, should be destroyed. And regarding sentence number two, my suggestion is replace term capacity to capability. It's more better from military understanding. Thank you very much.

Let me thank the distinguished representative of Ukraine for commenting on the text proposed. Next on my list is Pakistan, to be followed by the Asia Pacific Institute, the ICRC, and Israel. But let me first pass the floor to the distinguished representative of Pakistan. Sir, you have the floor.

Thank you, Chair. Good morning, dear colleagues. begin with, we would just like to echo the remarks of previous speakers in thanking you and your team for coming up with alternative formulations. And before sharing our comments on this, we would also like to highlight our support to the approach that you have taken. Rather than proposing changes throughout the text in these two boxes in a significant manner, which would have been very difficult to follow, I think you have chosen those strategic places in the text which could help bring everyone closer to a point of convergence. So that approach is very much appreciated, and we would encourage you and your team to follow the same approach throughout the course of this rolling text. Regarding the changes that you have proposed, I think that helps. add clarity in the text. The phrase which says within the scope of the application of the CCW and then in the second paragraph for the purposes of these elements of an instrument, we consider it to be the same thing. And I think when we look at the existing protocols in the definition section, at least we find this phrase very often for the purpose of this protocol. But right now, we understand that we have not determined what this instrument would be. So we can simply refer to it for the purposes of this instrument or for the purposes of elements of this instrument, as you have done in paragraph two. And that could also become a chapeau of all three paragraphs or the first two paragraphs at least. With this understanding, we recognize that we remain open to the idea of further changing this language once we have made that determination what this instrument is going to be in the end. Then in paragraph one, we support the word--the use of word "can" instead of "enable the system." I think that's more clear. Regarding The retention of integrated combination, we still have questions, some have raised by the delegate of Austria as well, but we do understand, uh, the points that have been expressed by a few other countries on this, so for this, you have chosen to retain this, uh, we understand that logic, regarding paragraph two, um, As we outlined previously, we do feel that there is a need to define the word "lethal" given the concerns that have been expressed by a number of delegates in our discussions previously. And we can go along with the proposed formulation. But having listened to the delegations of Austria, Brazil, and Cuba as well, I think dropping the word "weapons" and simply retaining with the capacity, the lethal means, the capacity, that could also very clearly convey the essence of this paragraph. Thank you once again.

Let me thank the distinguished representative of Pakistan for sharing his views and making comments. Next on my list is the Asia Pacific Institute to be followed by the ICRC, but let me first pass the floor to the distinguished, um, representative of the Asia Pacific Institute. You have the floor. Uh.

Thank you, Chair. Uh, we appreciate your efforts to take us closer to a text that could meet consensus and the opportunity to comment on this revised version. As you requested, we'll limit our comments to box one for now. The characterization of laws in paragraph one is now qualified by the phrase within the scope of the application of the CCW. The scope of application of a treaty, including the CCW, has multiple aspects. There's a substantive, personal, temporal, and geographic scope of application. The text of the CCW uses this concept in a particular way. Article 1 of the convention is titled Scope of Application, and that article makes the CCW and its protocols applicable in armed conflict. In accordance with amended Article 1, this encompasses international and non-international armed conflicts. In light of this language of the convention itself, paragraph 1 of the revised rolling text effectively means, in an armed conflict, a lethal autonomous weapon system can be characterized as, and so on. Accordingly, the new language of paragraph 1 does not seem to connect the characterization of laws to the objectives of the Convention, nor does it confine this characterization to the discussions occurring within the GGE. Yet, this appears to be the motivation for this revision. So we would suggest replacing the phrase "within the scope of application of the CCW" In its place, one option would be simply to use the formulation "for the purposes of the CCW". An alternative would be to use a phrase that has been used by this group before, namely "in the context of the objectives and the purposes of the Convention". Turning now to paragraph 2, which contains an interesting proposal that could address the concerns that many delegations have about the term "lethal". But there seems to be a slight grammatical issue. The word "lethal" is an adjective, so it cannot itself mean a weapon. It can only characterize a weapon. Therefore, the word "weapon" should be inserted after the word "lethal". The paragraph would then read, "For purposes of these elements of an instrument, lethal weapons mean weapons with the capacity to" and so on. Thank you, Chair.

Let me thank the representative of the Asia Pacific Institute for his comments and text suggestions. Next on my list is the ICRC, to be followed by Israel, Ireland, Singapore. But first, let me pass the floor to the distinguished representative of the ICRC. You have the floor.

Thank you, Chair. And thank you for the hard work from you and your team turning this around to have a revised text this morning. We generally support the changes you've made, and we very much agree with your comments this morning that the exercise at this point should be about formulating a broad characterisation, noting that this doesn't mean all systems will be prohibited, and it also doesn't foreclose the possibility of excluding certain systems later, which is already foreshadowed in paragraph 3. On paragraph 2, we echo the points made by Austria just now that the ideal resolution, in our view, still is deleting the word "lethal" entirely. But if it must be retained, we think the compromise clarification of "lethal" here is appropriate in that it captures not only the capacity to kill people but also the capacity to injure. And this is crucial because IHL protects combatants, for instance, against superfluous injury or unnecessary suffering, but of course, also protects civilians, not only against direct attacks that kill them, but also that may injure them. It's also imperative that the potential to damage or destroy objects is captured, as you've put on the screen, since, again, IHL applies equally to attacks against objects. This also aligns, we think, with the restrictions that the group has been discussing, including in box 3, paragraph 6, where the group has talked about military objectives by nature. These things are objects, so it's clear that the group has had in mind systems that can damage or destroy objects in the course of its discussions to date. On a drafting note, we do agree with Austria, Cuba and others that it would probably make sense to delete the tie to the weapon itself and to instead have the clarification relating only to the capacity. So that would mean deleting "weapons with" from the first line of this paragraph 2. Coming back on paragraph 1 and on the question of identify as a function, we reiterate the concerns that we raised on day one, and we also echo Austria, Brazil and others this morning in seeing the best way forward as having only select and engage as the critical functions. We believe that those terms have a more common understanding. including, for example, as stages in what are called joint targeting cycles of many militaries and organizations. In contrast, the function of identify does not seem to have such a common understanding, and so the term could introduce unhelpful ambiguity and arguably significantly narrow the characterization on some readings. In answer to how we see the term "identify," for us, it refers to the process of determining a general target profile or a list of potential targets, which could then be input to a system. And in our view, this process is often carried out by humans well ahead of the time that a system is activated or launched. In contrast, select then involves, after activation, a system essentially matching a specific person or object based on sensor data against that pre-programmed target profile or list. So again, we believe that a weapon system should fall within the scope of a characterisation, even if the identification of a type of target is carried out earlier by humans. On this basis, we continue to support deleting the word "identify," though we could see retaining the current formulation with "or" as an acceptable compromise here. Thank you, Chair.

Let me thank the distinguished representative of the ICRC for commenting on the text and making proposals with regard to how to deal with certain issues still outstanding. Next on my list is Israel, to be followed by Ireland, Singapore, and Switzerland. But let me first pass the floor to the distinguished representative of Israel. You have the floor.

Thank you so much, Chair, and good morning to all the colleagues. Thanks, Chair, for this interesting new proposal. We have been looking at this with the detail and working up to the last minute to have a detailed comment. That also means that this will be the first time that I read the text as a whole, so wish me luck. Concerning the rolling text in box 1, first, while this description attempts to address the meaning of lethality, unfortunately, this is done by giving this term an extremely wide meaning, making the practical implication of this section almost meaningless. This is due to the focus of this description on the capability of the weapon rather than on its design. As we know, IHL rules that -- sorry. As we know, IHL that regular weapons, including in the CCW framework, typically differentiate between weapons on the basis of their design. Indeed, almost any means of warfare and even physical objects which were not meant to be a weapon can hypothetically be used to cause death or injury. For this reason, the crucial element is the weapon's design. Therefore, we cannot accept the suggested terminology. We will replace the wording "with the capacity" with the wording "designed". Along other minor edits, the text would read, "A lethal weapon system means a weapon system designed to kill or injure one or more persons." Regarding the rest of paragraph two, international law requires us to interpret terms in accordance with their ordinary meaning. In the ordinary meaning of the word, lethal relates to the loss of life. However, it is unclear from the current text whether weapons systems capable of only harming objects could be considered. This is problematic since it can cause uncertainty and lead to illogical interpretations and outcomes. For these reasons, and as an alternative, we support in replacing the text that address objects with a different text in a new sentence that was previously suggested by the delegation of the US. "The fact that a loss can be used in circumstances that do not result in death, such as to destroy objects or to cause injury, does not exclude it from this characterization." In addition, for similar reasons, we think that the word "can" in paragraph 1 should be changed to "design" in line with the common terminology in international law applicable to weapons. We also note that it's unclear whether the word "can" refers to the weapon system or to supporting components, which is unhelpful. We reiterate our request to change the word "integrated" to "integral" so it will be clear we are referring to a definable weapon system and not to external components of general purpose, which are not actually part of the system. In our view, those changes are much needed to ensure stability and certainty. The current text does not give enough consideration to the fact that the weapon systems often rely on a large amount of information and external features that are not inherently related to it. The way it is phrased now makes it very difficult to determine what is considered part of the system and what's not. In addition, as we said at the beginning of the week, the functions "identify", "select" and "engage" are cumulative rather than alternative. Accepting the suggestion to see them as an alternative will introduce to the Convention systems that are not the focus of our discussion, such as precision-guided munitions. We therefore think that the word "or" needs to be deleted. Finally, earlier this week, we made a few comments regarding box 1. However, For the sake of compromise and facilitate consensus, we will not reiter- reiterate them. We did, we did reiterate comments which are the most important for us. Thank you so much, sir.

Let me thank the distinguished representative of Israel for sharing the views of his delegation and for making proposals with regard to the text on the screen. Next on my list is Ireland, to be followed by Singapore. But let me first pass the floor to the distinguished representative of Ireland. So you have the floor.

Thank you, Mr. Chair. Thanks as well very much to you and your team for this updated section 1. Like others, we appreciate your approach in this revision by focusing on areas of possible convergence, and we believe that the changes overall strengthen this important section. For the reasons outlined by Austria, Brazil, and others regarding the term "identify," for us, the "and/or" is critical here. Like Brazil explains, we're not discussing a prohibition of this characterization, so avoiding the cumulative approach is necessary at this stage. We welcome the clarification by what we mean by "lethal." Indeed, as we said earlier this week, we are not sure what a non-lethal weapon is. So if we are to keep the term, it is important that we are clear by what we mean by it. And this second paragraph goes a long way in achieving this. We would just join others in suggesting to tie lethal to capabilities or capacities as opposed to weapons. If we start to try to define what lethal weapons are, we are confounding the approach in paragraph one. So we would suggest instead paragraph two to read, "For the purposes of these elements of an instrument, lethal means the capacity to kill or injure one or more persons and/or the capacity to damage or destroy objects. So this would keep the focus on capacity and not weapons, as others have suggested. And just to add that for us, including design might introduce another loophole here, which we would suggest to avoid at this stage, again, maintaining the broad scope for the elements of an instrument. Thank you, Mr. Chair.

Let me thank the distinguished representative of Ireland for sharing his views and making comments on the proposals and reacting to the proposals of other High Contracting Parties. Next on my list is Singapore, to be followed by Switzerland and the United Arab Emirates, but let me first pass the floor to the distinguished representative of Singapore. So you have the floor.

Thank you, Chair. My delegation wanted to really express our appreciation to you and your team for the quick turnaround of the text. Regarding paragraph 2, Singapore is open to the inclusion of a definition of lethality, which has the potential to provide further clarity to our discussion. This helps to bring the meeting closer to convergence on our earlier discussions on the retention of the term. With regard to the current formulation on what lethal means, An intuitive understanding of this word is that it causes death. This element of mortality is crucial in differentiating the word "lethal" from other adjectives such as the word "harmful". One possible change to improve clarity could be to end the sentence at "the capacity to kill", which could help bring the working definition more in line with the spirit of the word "lethal". To be clear, the paragraph now reads: For the purposes of these elements of an instrument, "lethal" means weapons with the capacity to kill. In addition, my delegation is of the view that the current formulation with "capacity" in the paragraph is vague and potentially overly expansive. It could risk giving the impression that death and injury are comparable and permitting a very wide range of systems to qualify as lethal, such as tasers, pepper spray, high-pressure water cannons. So one possible option to improve clarity is to replace the word "capacity to" with "designed to" or "intended to." We are also open to the Asia Pacific Institute's grammatical suggestions. The sentence, therefore, now reads, "Lethal weapons means weapons that are designed to kill," or "Lethal weapons means weapons that are intended to kill." This could help make clear that loss is intended for the purposes of causing death. Singapore is open to further discussions with other delegations on how to improve this working definition of lethality. I thank you, Chair.

Let me thank the distinguished representative of Singapore for sharing the views of his delegation and making proposals. Next on my list is Switzerland, to be followed by the United Arab Emirates, the UK and the Russian Federation. But first, Switzerland, you have the floor.

Thank you. First, we want to thank you for the diligence with which you have discharged your duties and for having swiftly submitted to our consideration an updated version of the rolling text. We also commend your approach to move forward iteratively and in a targeted fashion when it comes to the way the text evolves. On para 1, in box 1, we would make the three following remarks. First, characterization that the GGE is working on only applies to the CCW because the contracting parties to our convention don't have the capacity to speak on behalf of other bodies or organisms. However, it is useful to clarify that state of facts if that facilitates the progress of our work and if it guarantees that this doesn't generate any confusion when it comes to the scope, especially when it comes to characterization. Therefore, we could support the addition at the start of this paragraph. Secondly, we welcome the modifications to the wording, enabling systems, enable system, and use the simple word "can". We can support that. Like other delegations have said this week, or this this morning, it's important for us to include systems that have the capacity to function autonomously within the characterization. As a priority in that characterisation, what we need to cover is the capacity of laws to operate autonomously in certain functions, and not the way in which the system operates. The new wording makes that aspect clearer, and we want to support it. Next. We think that a more in-depth discussion based on tangible examples is necessary to better understand the various approaches or understandings of the term "identify". That relates also to the terms "select" and "engage", as you yourself noted. For our part, we see a risk of the term "identify" If it is associated with a cumulative approach, it could significantly reduce the scope of the characterisation, and it might exclude systems that we think should be covered in our work. In that respect, we think that the deletion of the term "identify is an option we might envisage, indeed, we might desire, either since the deletion would allow us to make progress without doing any damage, for the following reasons. Selection means, by its definition, prior identification. The act of selecting a target intrinsically includes a form of identification. So, we think A system cannot select a target without first distinguish it from what is not being target. In other words, it's identifying it. And so therefore it's conceptually covered by the term select. Furthermore, the term identify that could lead to much broader interpretations or the inclusion of elements legally wider elements like recognition or general knowledge of the situation, and these aren't necessarily related to the functions of autonomy that lie at the heart of our considerations. Therefore, keeping the term "identify" could create confusion, meaning that all systems with capacity to identify could fall within that characterization, even if they don't select or engage targets autonomously. Finally, by placing emphasis on select and engage, the characterization targets specifically the functions of essential autonomy, which raise legal, ethical and operational concerns. However, whereas identify could introduce pointless complexity without really adding value. Therefore, leaving out the term "identify" would not weaken the characterisation. On the contrary, it would make things clearer and make the wording more targeted when it comes to the key elements of the use of the weapon system. The new paragraph 2 clarifying the notion of lethality, then. Like others, our preference would have been to delete that term in the first paragraph of the characterization. However, if that approach doesn't reach consensus, we would support your efforts to find a solution that would enable us to keep a characterization of the work whose scope remains sufficiently broad and to remember that the characterization is not the place where the means of excluding systems a priori, it should, that should be carried out in the following cases where the various types of uses and the various kinds of considerations of a legal, humanitarian, and other nature should determine the prohibitions and regulations. Your proposal clearly indicates what should be at the heart of our deliberations, in other words, weapons that have a capacity to be that weapons that have a capacity to apply force should be included in our work, and moreover stressing that this does not include other methods of war or weapons of another nature, such as cybernetic weaponry. We remain flexible when it comes to the exact wording of that paragraph, but we deem it important to include the considerations that I have just laid out. I thank you, chair. Let.

Me thank the distinguished representative of Switzerland, especially for starting the debate on on identifying. I think it's very helpful and some of y'all's already have also shared their views and I would really also invite other delegations to continue this discussion apart from giving us feedback of course with regard to the proposed changes. Let me now pass the floor to the distinguished representative of the United Arab Emirates to be followed by the UK.

The R thank you chairperson I thank you for your efforts on this wording we do not seek any radical changes but we have some comments to share on paragraph one we find that it would be appropriate to amend "without intervention by a human user" to include "without human control or control of a human user". In the second paragraph, we have no comment or objection. as regards using the term "lethal". But on the reference to weapons with the capacity to kill or injure, we propose replacing that by "weapons that have the capacity to kill or to cause injuries". And then coming to weapons that can be used to damage or destroy objects, we believe that that is compatible with international humanitarian law, which distinguishes between legitimate and non-legitimate targets in paragraph 3. Without prejudice to any future understanding, here we could say that the description could be developed and enriched in the future. And we would like to add a final sentence to paragraph 3, saying certain types of systems, depending on the degree of human control and the operational impacts. because that covers not only technical considerations but also human control. Thank you.

Let me thank the distinguished representative of the United Arab Emirates for commenting and providing alternative text. Next on my list is the UK, to be followed by Russian Federation, Germany, Australia and Belarus, but let me now pass the floor first to the distinguished representative of the United Kingdom. You have the floor.

Thank you, Mr. Chair, and thank you for all you and your team's efforts of turning this new text around relatively quickly.

Taking box one, this new text is acceptable to the UK.

With box two, as we said on Monday.

We recognize that there are many states who desire some sort of explanation of lethality and in that sense the explanation here is okay in principle with the UK but we are conscious that there will be many other delegations who will wish to discuss this further and the UK stands ready to offer its legal expertise in those discussions.

Thank you Mr.

Chair. would like to thank the distinguished representative of the UK for his reactions and for his offer to further continue an expert discussion. Next on my list is Russian Federation to be followed by Germany, but let me first give the floor to the distinguished representative of the Russian Federation. You have the floor.

Thank you, Chairperson. At the outset, we would like to echo the thanks express to you and your team for preparing this updated text of boxes 1 and 2 distributed yesterday. We regard this wording as a step in the right direction. It makes it possible to make progress towards a rapprochement of our positions among delegations. The Russian Federation welcomes this effort and confirms its readiness to participate constructively and substantively in this work in order to seek a consensus recommendation. To be specific on the paragraphs that we have in box 1, We would like to welcome the addition at the start of paragraph 1 to say that this definition is to be considered within the scope of the application of the CCW. Moreover, we like a number of other colleagues before us, would like to support the inclusion of the word "can" before the list of the so-called critical functions. We find this addition to add more clarity to the wording. However, we believe it is justified to have a reference to the fact that technological components, which we're talking about here in paragraph 1, have an auxiliary function. And it would be entirely justified to use the adjective auxiliary before them. Furthermore, we would like to confirm our position on the end of this paragraph. We believe that saying "in the execution of these tasks" is superfluous and that could be deleted. Since for us it is entirely clear when we're talking about the capacity of the weapons systems to carry out those functions that are listed in paragraph one without intervention by a human user, then we are talking about the stage of execution of those tasks. Moving to paragraph 2, chairperson, we welcome the attempt to achieve a unified understanding in regard the characteristics of lethality. That's fully in line with our approach, that determining this characteristic is necessary to provide a comprehensive nature to this working definition. However, as we see it, it would be justified not to take this characteristic out into a separate paragraph, but to try to integrate that into the definition, the working definition of lethal autonomous weapons systems. Since this characteristic is no less important than that of the autonomous nature, and it, in factual terms, is directly connected with autonomy, with that characteristic. So we would prefer to combine paragraphs 1 and 2 and try to have the characteristic of lethality included in the working definition. On the specific wording, of the characteristic of lethality, this, to a large degree, is in line with your own suggestions, Mr. Chairperson. However, in our view, it would be more justified clearly to set out the concept of causing harm to health or causing death and causing harm or destruction to objects. They should be separated more. We take the view that in terms of lethality, above all we're talking about causing harm to health or causing death. And as a compromise, taking into account what other delegations have suggested, we believe it would be possible to have a reference to damage or destruction of infrastructure objects as an accompanying effect where harm is caused to health or death is caused. And therefore the specific wording could consist of causing harm to health or causing death. which could be accompanied by damage or destruction to infrastructure objects. In English, it could sound like this:.

"Cause harm to health or kill that may be accompanied by damage or destruction of objects".

And this sentence could indeed become the end of the first paragraph. It would follow the verb "can" and would cover the fundamental characteristics or functions that this weapons system possesses. And coming to what is now paragraph 3, we would like to confirm our position on the need to include a clear understanding that this working definition does not cover separate types of weapons. We reiterate that we agree that at this stage there is no need to determine which types of weapons are not included within this working definition. But there is a need to indicate clearly that separate types will be identified and will not fall under the scope of this definition. I think overall we agree with this, namely that this definition will not be universal, that there will be certain exclusions. And so as we see it, it's important to reflect that understanding at this stage and do it clearly. In conclusion, we would like to support what has been mentioned by Australia, I'm sorry, Austria, regarding the fact that these paragraphs and overall the working definition meet the goal of functioning of the GGE. we have this concept within the chapeau, this covers the whole of the chairperson's document, but we believe that it would be justified were we to include concept specifically within section 1 in order to demonstrate clearly that the paragraphs within the box meet the objectives of the activity of the GGE. Thank you.

Let me thank the distinguished representative of the Russian Federation for giving us reaction and comments and suggestions with regard to the text put on the screen. Next on my list is Germany, to be followed by Australia, Belarus and the US. But let me first pass the floor to the distinguished representative of Germany. You have the floor, Madam.

Mr. Chair, Germany has studied the new version of Box 1, and we welcome your suggestions, bringing our discussions further forward. We support the new version of Paragon, however, we suggest to insert once activated before identify and/or select and engage. In our view, the temporal dimension would also be more precisely circumscribed if a further is inserted before intervention so that it reads without further intervention. We understand the arguments made by Austria, Brazil, Switzerland, and the ICRC suggesting to use "select and engage" instead of "identify and/or select and engage." We will.

Look further into this issue.

The new para 2 contains important aspects of our discussion on lethality, and we hope that it can find agreement in the room. We like the idea to link lethal to a weapon, ensuring close alignment to the GGE's mandate. In terms of understandability, the sentence might become clearer if we put "lethal" refers to weapons, for example, instead of "lethal" means weapons.

Thank you.

Let me thank the distinguished representative of Germany for commenting on the text provided. Next on my list is Australia, to be followed by Belarus, but let me first pass the floor to the distinguished representative of Australia. You have the floor.

Thank you, Chair, for your hard work and that of your team in updating the rolling text. Starting with para 1, Australia can accept the insertion of the phrase "within the scope and application of the CCW" as consistent with the GGE mandate and the CCW, noting that it was important for some delegations to be explicit in this respect. Likewise, we can support "can" in paragraph 1. This is streamlined language. The now view appropriately captures the sorts of weapons systems that are the subject of this group's discussions. That said, we would suggest a comma at the end of "technological components" just to ensure that the capacity to identify, select and engage can relate to both the technological components and the integrated combination phrase. Turning now to the definition of "lethality", with thanks again for your work on this. Chair, from our perspective, and that echoing the view that has been raised by others, lethality is integral to the mandate of this GGE. Our preference therefore remains the inclusion of a reference to lethal force in paragraph 1 rather than this definition, though we were interested in Singapore's proposed solution. For completeness, we have heard the interventions of others this morning and also wished to correct the grammatical issue in the current drafting. The Chair could take up the idea of APLS of inserting "weapons" after "lethal", or you could delete the words "weapons" with like the ICRC suggested, so that the term lethal more clearly refers to the capacity and not the weapon. We do look forward to working with delegations to achieve a consensus outcome on the issue of lethality. Thank you.

Let me thank the distinguished representative of Australia for providing comments and also for offering the availability of her delegation for further discussions on this topic. Next on my list is Belarus, to be followed by the US. But let me first pass the floor to the distinguished representative of Belarus.

So you have the floor.

Thank you, Chairperson. At the outset, we wish to express thanks for your efforts and for organising this effective work on the text, and indeed for the text itself. Briefly, some comments on paragraph 1, box 1. As we see it here, it is a very important provision, but it has essentially two approaches within it. We are seeing a connection with the CCW, which is a complex system related to the Geneva conventions and optional additional protocols, so international humanitarian law. We would point to the fact that the definition of lethal autonomous weapons systems involves a certain loop. There are components of weapons and technological components, and that is useful for the purposes of control, arms control and disarmament, but not very convenient, perhaps, for IHL, particularly as regards the use of the Additional Protocol II to the Geneva Conventions, which talk about irregular armed groups which could use elementary components. and connecting these with computers and operating systems, they can use these for destruction and for striking targets in that connection. This is not a criticism, but it's a proposal to envisage having two approaches, one for the purposes of arms control and the other to ensure the application of IHL, which must be broader, And if we're talking about the IHL approach, then it's more logical to exclude in the definition the words "weapons and technological" and leave only one or more components, which could be any. They could be widely available items that can be bought in shops and still they can lead to, they can produce little autonomous weapons systems. Thank you.

Let me thank the distinguished representative of Belarus for his comments. Next on my list is the US. Sir, you have the floor.

Thank you, Mr. Chairman. I just wanted to start by echoing comments that have been made by, I think, every delegation that's intervened so far this morning, which is to say that the substantial work that you and your team have done to present a new version have really put us in a good place to move this discussion forward. And we think that what you've presented us provides a very good basis for discussion this morning, as we've all observed. I think I would just like to start by making a simple point, which is that two of the key issues that we're discussing this morning have to do with with the letter L in laws, which is lethal, and the letter W, which actually is significant in a couple of different acronyms, both laws, it appears in laws, but it also appears in CCW, the W in CCW is weapon. And so we do welcome the focus that we've had this morning on these two terms, but the reason that And I don't need to remind people of these acronyms, but the reason that I'm stressing that these are terms that fall within the scope of the acronym is really to make a couple of points. First, that with respect to the word lethal, I think it's really important to understand as a starting point that lethal is in the mandate that we have. We welcome this discussion, but we don't really view this as a discussion that's attempting to achieve some sort of compromise about whether or not to include lethal, because we think that we are directed as a group of governmental experts to follow the mandate that we've been given by the high contracting parties, and it's pretty clear to our delegation that lethality is clearly within the scope of what we're trying to achieve. So what we very much understand is that we're in the process of trying to understand how best to clarify and to explain what we mean by that concept. We've listened very, very carefully to the discussion this morning, and And we're hearing a lot of very good ideas. And I'm not sure we are ready to decide that any one of these is the best possible approach because we think there are merits and advantages to several of the approaches that we've heard this morning. And so just to recap what we've heard and the way we've heard it. Well, actually, maybe to start with one that we haven't heard, but just to maybe put it on the table as maybe a simple way to think about this that could avoid a lot of the more difficult follow-on issues that we have with going to great lengths to provide additional definition. A simple solution could be to add the word lethal before the first reference to weapons and leave it at that. I'm not sure if we could agree to include that word at this point in this paragraph that we would need a lot of the additional work to provide these clarifications. So just for the sake of completeness in terms of identifying options, I think I would just want to remind people that that is an option for this group. Second, there have been a number of delegations that have come back to the initial proposal that we had a couple of days ago to address the question of lethal. I think both the delegation of Singapore came back to the idea that we had proposed, and I think other delegations did as well. And so just to remind folks that we did have an alternative proposal that we had identified yesterday. We did want to acknowledge an important point that the delegation of the Russian Federation made that I think has become clearer in this morning's discussion, that one of the risks of including a paragraph one and a paragraph two when we're trying to explain what we mean by the concept of lethality is that we potentially have created an artificial division between the characterization and what we mean by lethality, when in the view of our delegation, lethality is among the essential characteristics. So I think as a general proposition, we agree that combining these two paragraphs makes sense. And while we might have suggestions about particular wording, the idea of including the clause that the Russian Federation identified in the position that the Russian Federation identified also strikes us as an idea that we should continue to keep on the table. And then to come back to a suggestion that was made by the Asia Pacific Institute and again the delegation of Singapore, and this gets to my W in laws and CCW. I think for our delegation, having heard the discussion, a lot of delegations have suggested that we can delete the reference to weapon in what is now sort of the paragraph to approach that is in your current proposal, Chair. But for our delegation, I think the discussion highlights the need to emphasize the connection to weapons. And so we would support the suggestions made by the Asia Pacific Institute, which I think were echoed by Singapore, but I may be misspeaking here, to incorporate a reference to weapons more clearly in paragraph two. I think their suggestion was to say lethal weapons means weapons, et cetera. So we liked that suggestion. To pick up on some points that our distinguished delegate from Australia made, and here this may be a debate that's unique among the native English speakers. We did agree that including the comma after the word "components" in paragraph 1 is an important grammatical addition because, like the delegation of Australia, we do want to stress that what we're talking about relates to both components of this combination. Another way that we think could also make that clear would be to change that to the word "which." But I think the essential idea of including the comma there is something that we think makes a lot of sense. The reference to human user, I think, continues to create some concerns for our delegation. I think the main concern that we have -- well, we have two concerns. We continue to think "operator" is simply the better way to refer to the particular user that we have in mind, both because as a matter of our practice, that's how we would view the right human being to be identified. But we also continue to have the concern that if we add human user here, we are actually narrowing the scope of the definition. And just to remind, we've spent many years focused on the sunrise, you know, terminology and the fact that we are looking at a life cycle. And a human user could theoretically refer to a human user at an earlier stage in that life cycle, or in an earlier place in the sunrise graph. And we think it's important, as several other delegations have emphasized, to bring with this characterization the idea of the temporal aspect of the intervention that we're talking about. And we think "operator" continues to best capture that. Chair, I think the last observation that we would make is that we are listening very carefully to the debate about identify, and like many other delegations, we do think that this is an issue that would greatly benefit by more concrete examples of what it is that we're talking about, both in terms of what delegations are concerned will be excluded and what delegations are concerned will be included. And that's the value of this body, that we have the technological, the military, the legal expertise to have that much more detailed technical discussion. And so we think that would be helpful to do. I think for our delegations, from our delegation's perspective, like many other delegations, We initially were very attracted to references to select and engage. I think we continue to be attracted to that, but we understand the complexity of this particular issue and are very, very eager to continue to work with everyone to try to find a way forward. Thank you.

Let me thank the distinguished representative of the United States for sharing the views of his delegation and for commenting on the text proposals, as well as offering to participate in a discussion with regard to identify. Next on my list is China to be followed by the Republic of Korea and Australia. Is this an old blackboard? Okay, good. Then it is China to be followed by the Republic of Korea, but let me first pass the floor to the distinguished representative of China.

Mr. Chairperson, China would like to join other delegates to thank you and your team for your hard work and for the updated rolling text. China stands ready to work with all parties to seek more convergence and consensus. Specifically, regarding paragraph one, the working characterization of the laws, China believes that the laws should include all functions, all three functions of identify, select and engage. Therefore, we suggest deleting the word "or". As for paragraph 2, we thank the Chairperson for adding a description on lethality. China notes that currently the word "capacity" is used in the text. China agrees with Singapore and other delegates that the term "capacity" may lead to an overbroad scope. Therefore, China suggests changing it into "design tool." As for paragraph 3, in the first round of discussion, China suggested including an exclusion clause and an open list of exclusion. to make clear which systems do not belong to laws. China still believes this is a valid approach and we would like to have further discussion on this issue. Thank you, Mr. Chairperson.

Let me thank the distinguished representative of China for commenting on the text proposed and for offering alternative language. Next on my list is the Republic of Korea. You have the floor.

Thank you, Mr. Chair. My delegation joins others in expressing our appreciation for your leadership and the hard work of your team in guiding the group's progress. Now, some delegations suggest that the working characterization should be as broad as possible, given that we are not discussing prohibition or regulation. However, my delegation believes and emphasizes that our mandate define and delimit all our discussions and our work. The working characterization is meant to serve our work and cannot be indefinitely expanded, as this would contradict the very notion of working. Regarding paragraph 2, my delegation wishes to offer our preliminary view. In our view, it is concerning to us that the new formulation deviates further from the ordinary meaning of "little", bringing us closer to creating a new vocabulary. In this regard, we believe that the original formulation proposed by the US focusing better on the ordinary meaning of the lethal, which pertains to causing death, is better. Or the simple fix of putting lethal in front of weapons in the first paragraph proposed today can be a good and straightforward solution. Additionally, my delegation is open to considering the suggestions made by Israel and Singapore this morning, as they may help align the term more closely with its actual meaning. And I would like to reemphasize what I said on the first day of our meeting. The word lethal should be included in our characterization, not merely because it happens to appear in our mandate, but because the mandate and the original purpose of this group are to discuss a distinct category of weapon systems. That is why we have lethal in our mandate in the first place. Therefore, we cannot simply retain the word while defining it in a way that contradicts its actual meaning. This, again, does not really contribute to the notion of working, but rather it is counterproductive. Thank you.

Let me thank the distinguished representative of the Republic of Korea for commenting on the text and making proposals. Next on my list is France, to be followed by Luxembourg. Let me first pass the floor to the distinguished representative of France madam you have the floor.

Thank you chairman first I want to thank you for your efforts and that of those of your team to move the text forward on the new Patta 2 we could support the substance but We continue to believe that it's not up to us to define the word lethal or the notion of lethality, but rather to clarify the characteristics of lethal autonomous weapon systems. We remain flexible in terms of how we can do that without modifying the first part of the sentence, lethal means weapons. And other delegations have said before me, we also see a grammatical problem with that expression. Or by deleting it and merging the rest of the sentence into para 1, then it would read...

"After in the execution of these.

Tasks..." We'd add a comma. followed by...

...With the capacity to kill.

Or injure one or more persons, which includes weapons that can be used to damage or destroy objects.

Finally, our delegation would like to reiterate two points which are still lacking in the definition. We still think that the definition must refer to the notion of activation. and to defining the framework of use prior to the autonomy stage. In that respect, we support adding the expression "once activated", as suggested by Germany. That would allow us to identify the moment when the human determines the framework of utilisation. We'd also add the expression "further" because without that term, the definition would only apply to completely autonomous systems. Alternatively, we'd like to maintain the expression "in the execution of these tasks". which allows us to delimit autonomy of execution of tasks, which presupposes an activation or prior programming. Finally, we continue to support the expression "without further human intervention within the chain of command and control", specifically to avoid excessively expanding the scope or limiting it to the military and operational spheres. Thank you. Let.

Me thank the distinguished representative of France for commenting on the text and making proposals. Next on my list is Luxembourg. Let me now pass the floor to the distinguished representative of Luxembourg. You have the floor.

Thank you, Chair, for giving us the floor. We welcome your next text proposals and the step-by-step approach you are taking. Regarding the first paragraph, we want to add a voice to the delegations that have proposed omitting the mention of, of identify, as in our understanding, identification is interesting, intersectly and inherently linked to the selection process. However, in a constructive spirit, we remain flexible on this point. Regarding the second paragraph, while Luxembourg's national position on autonomous weapons systems does not include the notion of lethal, we recognize the importance that some delegations attach to it, and as the United States have underlined, it is an integral part of our current mandate. Therefore, we are open to keeping lethal in the working definition. but support an extensive interpretation of lethality to avoid narrowing the scope of the characterization. We believe the definition you have put forward in paragraph 2 fulfills this purpose. However, as mentioned by many other delegations before us, we would prefer to characterize lethal in relation to the capacity to kill or injure rather than linking it directly to weapons as it is currently formulated. I thank you, Chair.

Let me thank the distinguished representative of Luxembourg for commenting on the text and also for showing flexibility with regard to the proposals made so far. Next on my list is Sweden. Let me pass the floor to the distinguished representative of Sweden. You have the floor.

Thank you, Chair. We welcome this text and commend all your efforts to advance our work. We also appreciate the current discussion with constructive comments from all delegations. We prefer para 1 as it is in regard to identify and/or select and engage. We can also accept an addition of once activated proposed by Germany, and we support the comma proposed by Australia. As regards lethal, we prefer to remove lethal in para 1 and also remove para 2 as we do not see a need to define lethality. But we hear that this aspect is important to several delegations and could therefore accept para 2 as a compromise. We could also accept replacing capacity with capability or are designed to. We could also replace lethal with loss or with a reference to para 1. However, having listened carefully to the discussions this morning, we would also like to add another possibility as a helpful suggestion. We wonder if it would be helpful for states that are hesitant to para 2 to reformulate in the following way. For the purpose of characterizing loss in these elements of an instrument, comma, "Engage a target" means using force to kill or injure one or more persons, damage or destroy objects. I will send the sentence to the Secretariat, of course. This would, in our view, mean that we refrain from defining the term "lethal" while maintaining the intended purpose of para 2. We are also open to other reformulations that maintain death and injury to persons and damage and destruction to objects, as we believe that it's military, economically, and legally sound to formulate elements for AWS that are designed to be used to target all types of lawfully military objectives. Thank you, Chair.

Let me thank the distinguished representative of Sweden for commenting on the text and making suggestions, and indeed, we would appreciate to receive them in writing. Sweden was the last on my speakers list. Are there any further requests for the floor? Oh, I see. So I do apologize. You're a bit out of sight. So let me now pass the floor to the distinguished representative of Mexico. You have the.

Thank you very much, Mr. Chairman. First of all, I would like to congratulate you and your team for your sterling work and the positive way in which you are steering our discussions. With respect to box 1, my delegation agrees with the previous speakers that the new proposal is positive. We have a few comments similar to those already made. With respect to para. 1, the phrase proposed by the Asia-Pacific Institute, which is broadly used in prior discussions of the GGE, would be a better way of addressing that paragraph, referring to the principles and objectives of the CCW. We also support the use of the phrase "can enable". I think that helps us better send to the definition on the human element. Finally, on the term "identify", we agree with the explanations given by Austria, Brazil, ICR, Switzerland and others. that its use does introduce ambiguity into the paragraph, does not address the critical functions of concern that raise ethical and legal questions, and which are implicitly covered within function and selection. On para 2, my delegation would prefer leaving out the word lethal as we have discussed in previous years in the GGE. We understand that certain countries find it important, and therefore that definition could be useful. It is important that the definition be aligned with the CCW and with the norms of IHL, which include causing excessive injury and having indiscriminate indiscriminate generic effect. In that respect, that vision is aligned with the broad vision that is presented here. Finally, we support the improvements to... We could have lethal weapons or we could just leave it only at the beginning or just cover capacities. We could do it one of those ways. Thank you.

Let me thank the distinguished representative of Mexico for commenting on the text and on proposals made by other delegations. I have four further, five further requests for the floor. I have Belgium, Austria, Norway, Japan, and Bulgaria. But let me first pass the floor to the distinguished representative of Belgium. You have the floor.

Thank you, Mr. Chair. My delegation wishes to echo some of the other delegations that expressed their appreciation for the approach taken by you. Regarding paragraph one, we welcome the proposed change to can identify as having the capacity is not the same as actually doing something. We also welcome further discussions on the interpretation of the wording and integrated combination regarding the new paragraph two. If retained, we welcome your proposal to clarify the meaning of "lethality" in the context of these discussions. We appreciate the broad interpretation given to the term "lethality", which should be read in conjunction with the two-tier approach contained further in the text, distinguishing between those laws that are to be permitted and those that are not. We would like to join other delegations in expressing our preference for replacing the capacity to kill with the wording "designed to kill". Furthermore, we agree, grammatically speaking, with the remarks made on the wording "lethal means weapons". In our view, lethality should best be described as an adjective. Furthermore, we wish to suggest replacing "injure" by "harm". In our view, "harm" would better encompass all possible intended effects. These may, for example, be to cause physical or psychological illness, something that would be something that would be covered by the term that would not be covered by the term injure. Regarding what is now paragraph three, in our view, paragraph one should always read in conjunction, should be read in conjunction with paragraph three. I thank you.

Let me thank the distinguished representative of Belgium for commenting on the text and also on proposals made by other delegations. Next on my list is Austria to be followed by Norway. But let me first pass the floor to the distinguished representative of Austria.

Thank you very much and sorry for taking the floor a second time. We would like to focus on the two issues, the lethal and also some other elements that have been suggested this morning and have some quick reactions on what other delegations have said on these two points. We heard a lot of arguments that Liefel is part of our mandate and that it has kind of rooted in our history, but I think we should also recall that this has not always been undisputed and this was to a large extent a compromise at the very beginning to include this word, and we are now coming back to the original discussion. Probably it was dormant now for many years, and we are now coming back to something that our predecessors wanted to put aside. This is probably natural, but at the same time, we also need to stress that just because it is a part of the title of our working group does not mean that we will not have a discussion about it. And we also noted that there was a certain kind of renaissance of this term because there was only a small number of delegations who thought that this makes even sense. And we were surprised about the many positions that have changed this week, to be very honest. On this discussion, once again, obviously, there are non-lethal weapons, such as blinding laser weapons, and the protocols that we are having, there are rubber bullets have been mentioned, and water cannons, and other things. But this does not mean that all non-lethal weapons are not regulated under IHL. And as we all know, riot control agents are not allowed as a method of warfare. So this is not per se an idea that non-lethal weapons should not be included in IHL instruments, so this logic must be dismissed. The second question that we must pose to ourselves here is that what is a non-lethal autonomous weapon systems and for what purpose would it be used? And this is just in addition to the many arguments that have been made by the ICRC and others that the lethality is not a category in itself. in international law, and that therefore it would be a quite novel thing to do here. And we should be very aware here that we are going into new grounds here on how to interpret IHL and setting kind of a new barrier on weapon systems and actually a new category or even a new separation between weapons that should be non-lethal and lethal, going away from those a few examples that you already know, and putting it into the middle of warfare, just to set this out as well. This next point that I wanted to make with regard to lethal is-- and it has been brought up-- that IHL is broader. It protects civilians not only against death, but also against injury. And as some, I think, have already pointed out, also against unnecessary suffering. And here we have a little bit of a problem with our text because this is something that we also spell out in box three. So we would exclude a certain element in the characteristics and then we introduce it in box three. So there's also certain logical problems if we are too narrow in what we are doing here in the characterization. Then we also wanted to react to the idea introduced by Israel and Singapore to use the word to be designed. there are two problems. I mean, the first one is that we have the experience that this creates a loophole, that you can have a weapon systems that is not per se designed, um, to be, uh, lethal or to kill or injure, uh, but there are still possibilities and ways to use it. And this in itself is also already against to a very, strong idea and notion that we know from IHL, and it's a huge loophole that unfortunately is already in practice to some extent. So this is something we, from our side at least, want to avoid. Then the argument that this is something that is common, we did a quick check, and of course there are protocols of the CCW that use in their definition the design verb. But in most of the cases, it's an effect of the weapon. It's designed to cause permanent blindness in the case of the protocol on blinding laser weapons, or it's designed to set fire to objects in the case of protocol three. So it explains what the weapon does. Then, of course, there is the example of anti-personal mines, where we have this definition designed to kill, injure. But we have to be aware that this what you are talking about when we talk about anti-personal weapons is an anti-personal category. And if we go into this direction, we all in this room should be very aware that then we are excluding a large part and we are doing regulation on anti-personal weapons. And I think this is not the original intention. It is a large limiting of our work. And it is also changing its nature after 10 years. And this is something that we would not see in general a good faith effort to be, again, a little bit more honest about how we see this. There are other designed objects. I mean, mines are also defined as designed to be used under Amenda Protocol 2, but they, of course, they include the damage element, and this goes then probably in the direction of the proposal of the Chair, that if we, when we use this kind of element, then we are going into a very broad definition, but something that is very important. to keep in mind is that Amended Protocol II has very strong general provisions that apply not only to anti-personnel mines, but to all kind of objects that are regulated under Amended Protocol II, and they go far beyond what we are trying to set up here for the purpose of autonomous weapon systems. So there is a different quality even to those weapons that are using this kind of definition. So we would definitely have a strong reservation of using design 2 in any kind of lethal characterization. The idea to fuse the two paras, I think it defies to some extent the original idea of what the second para should do. It was not to expand the definition, but the idea was to explain what the word lethal means. And again, in our mind, this is not a para that necessarily has to stay here. So linking it up with the main definition and what we see as a working characterization would go against our position, I have to say. It's better to have it as an additional working assumption and not as part of the main characterization. I thank you.

Let me thank the distinguished representative of Austria for commenting on some of the previous proposals, um, made in the room. Next on my list is Norway to be followed by Japan and Bulgaria, but let me first pass the floor to the distinguished representative of Norway. You have the floor.

Thank you, Chair, and thank you for the work you've put into providing us with this new text, which really seems to be taking us in the right direction. For us, it's important that the characterization remains fairly broad and that it doesn't define laws too narrowly. As we see it, the kinds of measures and regulations that we have identified so far in this document could, in principle, apply to all weapons systems. They would just not be very relevant to systems that are not autonomous. The issue of human control and deactivation, for instance, simply doesn't arise in the context of, say, a gun. So, so we don't think that we risk much by having a definition that is somewhat broad, um, and, and the concept of autonomy anyway makes it abundantly clear that we're not talking about guns or other systems of that kind. And if down the line we identify limits and regulations where a broad characterization would provide a problem, uh, paragraph three makes it clear that we should come back and adapt the characterization accordingly. Um, and as Brazil pointed out, I think it is, of course, important also to remember that this characterization does not imply that all those weapon systems, uh, which fall within the characterization would be illegal. It simply defines the scope of, uh, what we are addressing in this prospective instrument. Against that backdrop, we would be in favor of simply deleting the characteristic that laws be able to identify targets autonomously and at the very least keep the alternative and/or formulation for now. In other words, that we not have a commutative approach where the system needs to both identify and select a target autonomously to be a laws. It can of course do both, which would still make it fall within the characterization, but it shouldn't have to. And it's not clear to us why a weapon that could only select and engage a target autonomously should be excluded from the scope of laws. And we're also not quite clear on the meaning of identifying this context. As we see it, narrowing the scope with a cumulative approach could make it too easy to circumvent the characterization. On the issue of lethality, we believe that a definition would be useful to build a common understanding of what this means in the context of laws. And at first glance, the definition that you have provided us with in the new paragraph 2 looks like something that we could go along with. It makes clear that laws encompass those weapon systems that are not only designed to kill humans, but weapon systems more broadly that could be used for this purpose or to have that effect, including those that can be used to damage objects. I think it was the Israeli delegation, maybe, who mentioned that in principle, any physical object can be used to kill and damage. And we agree that those kinds of objects are not the concern of this GGE. But this is where your definition is quite clever, because the last clause, which reads, "includes weapons that can be used to damage and destroy objects," it makes it clear that we're not talking about objects in general, but precisely about weapons. If we were to focus only on weapons that are designed to kill humans, we would effectively be characterizing laws only as anti-personnel systems. And as the ICRC succinctly pointed out, IHL addresses more than just that. So if the intention is indeed to narrow this characterization down to anti-personnel systems only, we really would like to understand the rationale for that better. On the specific wording of paragraph 2, we appreciate the point that the current text seems to suggest that lethal and weapons are synonymous, and that we think could be easily fixed, and many good suggestions have been put forward. One could be to simply say that lethal means that the weapon has the capacity to, and so forth. On the issue of whether we should explain, uh, lethal weapons as a, uh, in a separate paragraph or just as part of the characterization in paragraph one, uh, we're open to both solutions, uh, I think for our discussions now, it's maybe convenient to have it split and not sort of overburden the characterization in, in, in paragraph, uh, one, which, uh, also has some issues that we need to sort out. But we are open to combining them and Frans' suggestion to layout the meaning of lethality as a clause at the end of paragraph one could work for us. On paragraph one, we can support the replacement of enable the system with can, and we can also support the inclusion of within the scope of the application of the CCW at the beginning there. But we agree with, I think it was Pakistan who mentioned that maybe it might be better to write a scope of application of this instrument since we are proposing elements for an instrument. Last point regarding this issue of human user in paragraph 1, we wonder if the concerns that the US delegation expressed just now could be addressed by changing it to human operator. I mean, it might appear superfluous, but it at least helps to ensure emphasis on the human element. Thank you.

Let me thank the distinguished representative of Norway for commenting on the proposals made. Next on my list is Japan, to be followed by Bulgaria and Italy. But let me first pass the floor to the distinguished representative of Japan.

Thank you, Mr. Chair.

First, Japan would like to join others to thank you and your team for your leadership in guiding our discussion forward and providing us with a revised version of the rolling text. First, on lethality, We also think that it is part of the mandate and provides important focus of the work of this GGE as articulated by several delegations who took the floor before us. Our preference is to have a reference in paragraph 1, but we have carefully listened to other delegations and we are open to working on description. We also consider that temporal aspect is important.

Therefore, we support adding "once activated".

As proposed by Germany and supported by France. Finally, we can also support the proposal by France to add "within the chain of command and control".

At the end of paragraph 1.

Thank you, Mr. Chair.

Let me thank the distinguished representative of Japan for sharing the views of her delegation. Next on my list is Bulgaria, to be followed by Italy. India and Singapore, but first the distinguished representative of Bulgaria, you have the floor.

Uh, thank you, Chair. Uh, we welcome your and your team's work on, uh, Box 1, and we have a few comments. Uh, we support the introduction of, uh, CAN instead of that enables, enables the system. I guess we can support the native speakers, namely Australia and US on the comma after technological components. Still on para one, we would prefer the inclusion of further before intervention. And we can also support the proposal by Germany of the introduction of once activated. On LITO, we understand the challenge of coming up with a definition. And we can work with the current one, however, the formulations, the formulation looks a bit strange grammatically.

To us and the proposal.

Made by Apos and the one from Germany, even the one from Norway a few minutes.

Ago could work for us.

Thank you.

Let me thank the distinguished representative from Bulgaria for commenting on the text. Next on my list is Italy, to be followed by India and Singapore. Let me now pass the floor to the distinguished representative of Italy. Madam, you have the floor.

Thank you, Chair, for giving me the floor, and thank you for the hard work you and your team have conducted on this section of the rolling text. We know it's a very difficult task. Looking at the new proposal, we see that it does not favor the cumulative approach referred to the three critical functions of laws identify, select, and engage, stressing the interventions of different delegations over these days, most recently by the distinguished delegation of China. In our view, if one of the three phases is replaced by the intervention of the human operator, we are not dealing with the fully autonomous laws. We do not share the point shared by some distinguished delegation earlier that select conceptually includes identification. For us, these are two distinctive phases of the process. So we reiterate our preference with removing the word "or" from the proposed formulation. We instead support the inclusion of "can" prior to the list of critical functions, even though our preference would be for the wording as the capability to. Another relevant point for us is to add after activation and the further before intervention as stressed a few moments ago by the distinguished delegation of Germany. Finally, on lethality, in light of our mandate, we continue to be open to formulation of LIDL as we heard earlier this morning that places at the center the injury to humans excluding objects that are not connected, even indirectly, to the possibility of causing damage to human health and ultimately death. And I thank you, Chair.

Let me thank the distinguished representative of Italy for sharing her views and making comments. Next on my list is India, to be followed by Singapore. But let me first pass the floor to the distinguished representative of India.

Thank you, Chair, and a very good afternoon. At the outset, let me please thank you for providing us the revised text and all the hard work that has gone into it. At this point, we will restrict our comments to making textual suggestions because we believe that the focus of our discussions should be to reach consensus on the language of the possible elements of this instrument. Our comments are hence preliminary and also brief. We support the Israeli, Chinese and Singaporean interventions to introduce the element of design, which we feel is essential in the characterisation of laws as we are discussing it in the scope of this GGE. The addition of this element in the characterisation would facilitate the adoption of the text as it exists for my delegation. Secondly, Chair, we support those delegations that have stressed on the importance of endorsing the cumulative nature of the individual critical functions of identification, selection and engagement without any human intervention for these weapons to be characterized as laws for the purpose of this instrument. Therefore, we would support identify, select and engage a target in the text. Thirdly, we would also support the views expressed by the US and others that the mandate of this GGE is to discuss only lethal autonomous weapon systems. We would thus support language that respects this mandate. Finally, we would also support the German and other delegations who have proposed the insertion of "once activated" before "can identify, select, and engage a target" to clearly reflect the temporal aspect of autonomy under discussion. I thank you, Chair.

Let me thank the distinguished representative of India for her intervention and for the comments and suggestions made with regard to the text we are discussing. Next on my list is Singapore. You have the floor.

Thank you, Chair, for giving us the floor again. On box one, paragraph one, my delegation notes your request for further explanations on these two issues. First, the difference between identify and select. Second, why the three components identify, select, and engage should be listed cumulatively. We will share our perspectives on both questions. On the first question, My delegation understands the word "identify" to be the ability to discern the specific type or characteristics of an object or target, whether it is a friend or foe, for instance. The word "select" refers to the ability to prioritise or choose a target based on rules of engagement, mission parameters, or mission objectives. From an operational standpoint, these two terms have distinct meanings and nuances. I can identify a target as a foe and decide not to prioritise it. In other words, I can identify a target without selecting it. Theoretically, you could also select a target without first identifying it, but that may neither be responsible nor logical. We therefore are not yet convinced by suggestions from some delegations to drop "identify". Moving to the second question on why the terms need to be listed cumulatively, my delegation wishes to provide some examples of weapon systems. An example of a weapon system that would identify and engage, but required a human to select, would be a fighter aircraft. Its onboard systems would be able to identify potential adversaries. However, a pilot is required to select a particular type particular target to attack, based on his tactics and priorities, following which a missile is launched to engage the adversary. Another example of a weapon system that could select and engage, but would require a human to identify a target, would be a Close-In Defensive Weapon System. After the human operator identifies actual threats based on parameters such as speed and trajectory, The defensive system would select and engage the targets depending on priorities, such as proximity of the threat, probability of hit. To recall, the rationale for having a working characterization is to establish a shared understanding of systems that would be considered laws. More importantly, to operationalize this characterization effectively and practically on the ground, it would have to be very clear and legible to capability developers and military operators of states. If the current formulation of and/or is retained, it could complicate the operationalization of this characterization, as it would appear to include weapon systems that would intuitively not be considered loss, as we have illustrated with those examples. Those examples intuitively are not the types of weapon systems that we are seeking to govern in the context of this meeting. So we hope that our explanation has helped to shed some clarity on the subject and we look forward to continued discussions on this. I thank you, Chair.

Let me thank the distinguished representative of Singapore indeed for, um, sharing some examples. Uh, and he was the last on my list. We have 10 more minutes. Yes, uh, Austria, you have to….

I want to thank the delegation of Singapore for bringing up examples, because I think this is the kind of idea that we need to move forward. And I think the very important point and maybe the kind of problem that we are currently facing is the room that some of us see some weapons, if we use the word identify, some weapons that, as you mentioned, intuitively and probably also by current practice, would not be considered an autonomous weapon systems would be included. But the same problem goes into the other direction, that there are also some weapon systems, if we have this cumulative approach that all three actions have to be included, that we would exclude some weapon systems that we would consider autonomous weapon systems intuitively in the other direction. So the problem goes both ways, and probably the issue is where are you coming from. Do you want to have a broad encompassing characterization that includes all the weapon systems that definitely are loss? But we also understand completely the other approach that you want not to have a definition that is just covering things that by nature are not loss. So the question is, how do we solve this? And I have no good solution at this point. But I just wanted to point out that it also goes into the other direction and that we are thinking, for example, on systems where the identification has been done at a very early stage of the process and where the autonomous weapon systems with a strong capacity of autonomy actually selects and engages a target. So the autonomy in the operation is given and everyone by intuition would consider this an autonomous weapon systems, and there are systems like this already, um, in place. Um, another example, and, and, and people have pointed out to me that it's not necessarily the best one, uh, would also be a swarm where the selecting is just one of the main, uh, capacities of, of, of, the distribution of targets and which unit in a swarm would select which target is one of the key capacities. But if there is a pre-identification of a large number of targets, then under the current, under the cumulative definition, a swarm would be included. But by intuition and also by its main capacities, and it would be an autonomous weapon systems. And as a second point, many of the concerns and issues that we try to address in the boxes two to five equally apply to those kind of weapon systems that do don't have an identify element, um, and to stress once again, there's also the question of distance. temporal, but also geographical, if the identify and the select are taken at completely different places and completely different points in time. So this is also something that is important. You pointed that the functional approach is in there with the integrated combination, but still this is something to also keep in mind. And the cumulative approach would exclude many systems that we would consider as autonomous weapon systems. very clearly and very intuitively. I thank you.

I have one more request for the floor. Let me pass the floor to the distinguished representative of Brazil. You have the floor, sir.

Thank you very much, Mr. Chair. And my delegation would also like to express its appreciation to the approach that was just demonstrated by our distinguished colleagues from Singapore. I think that, uh, what his intervention should be a model to the engagement that we need to have on this issue, uh, uh, in terms of, uh, adding clarity to the understanding of these three terms, identify, select, and engage, uh, uh, and, uh, adding real world examples of what is underpinning these concerns. Without this, I don't think that we can have a fruitful debate. So again, I would like to thank Singapore for this and to call on others to follow this approach. Regarding the points that were made, we would have the same concerns that Austria has expressed. These real-world examples are seen by my delegation on a different light. Closing weapons systems, as they usually called CWIS, in our mind, do identify, select, and engage. A human operator does not have to see a threat and activate it. Usually, these systems are used in a fully autonomous manner when in a context that requires them as a defensive measure. This is just an example of how we approach the idea of identification differently. The identification along the lines of what we heard would be the definition of the target profiles, which I think would be a design and development phase, not the target engagement phase. But again, I think that the risks of adopting these cumulative approach with the three terms was potently exemplified by Austria. when it comes both to loitering munitions in which the identification can be done by a human, but the selection and engagement can happen at a much later stage. And also when it comes to swarms in which different elements, different components of the swarm might do just one of these functions at a given time, but in a distributed fashion work as a coherent whole. So I think that these concerns are very much present and inform our attachment to a cumulative approach using only two critical functions, select and engage. As a food for thought, I would offer the possibility of removing a comma, which would, if that can be identified as a possible way out. It would be one of those proverbial multilateral moments in which removing a comma might do the trick. But removing the comma between select and engage might give the impression that we want, which is select and engage are part of the same process and they're not. And we remove the notion that identification and selection can be interchangeable. Because one point that our distinguished colleague from Singapore made, with which we concur, is that just identification and engagement would unnecessarily open the scope of this characterization, even to fighter aircraft, which I think is not what we want here. wrap it up, our delegation strongly prefers retaining only select and engage, but if we are to include identify, maybe removing the comma could be a way out. Thank you.

I see one more request for the floor. Ukraine, the distinguished representative of Ukraine, you have the floor.

Thank you, Mr. Chairman. I'm sorry, my view will be shortly. My remark after speech of representative of Australia regarding swarm technology. In this case, two element of swarm can be select the same target as example. But second study should be identify. It is target is free or already take to destroy or impact. In my opinion, first study should be select. Coma, then it should be identified. It's my last, uh, clarify. Thank you.

Thank you for, uh, again, showing your perspective on this, um, uh, on this very difficult discussion. Um, we have come at the end of our morning session. I would really like to thank you all for a very interactive debate, and indeed, as it was stated, um, at the end, we got some of the, um, some of the, also, uh, examples of, um, what is to be or not to be considered, uh, a loss, uh, in relation to the discussion on, um, uh, identify, select, and/or engage. I have the impression that we might still can do with some further discussions on this topic, so I'm a little bit hesitant already to close it. although we are now at the end of our list of speakers, but that doesn't mean that is the case again after the lunch break when people have had some time to reflect on it. And we have, of course, box two to consider where we have a few changes proposed, and we would like also to have your reactions on that. With that in mind, we will continue our discussions at 3 o'clock, and for the moment, this meeting stands adjourned.