8th Meeting - 1st Session Group of Governmental Experts on Lethal Autonomous Weapons Systems 2026

Excellencies, distinguished delegates, could I ask you to take your seats, then we can start. Good. Excellencies, distinguished delegates, let me start by thanking you all for this morning's discussion. I think it was most helpful. And because it was so helpful, we will continue it tomorrow, I think. And we are already working on some tweaks to the text on the basis of the discussion we had this morning. And I think there is further room for convergence, and I will try to harvest that tomorrow morning. Now, with regard to the rest of today, my intent is to have a second reading of box 3, and then we have a small break, and then at 7 o'clock again in room 11 for high contracting parties in the ICSC, we have informal consultations with regard to, I would say, to the Paris 6 and 7. unless something very strange happens and we all agree them at referendum this afternoon, but I don't think that will be the case. And as I stated earlier, I think that the issue of human involvement is something we have to pay extra attention to during this week. in order to close still existing divergences. Now, with regard to the suggestions, again, suggestions, this is not a new version of the wording text, the suggestions we sent out yesterday evening. With regard, I start with box 3 because In my view, the suggestion in box 2, para 4, is a result of a possible change in box 3. So for that reason, the logic is to start with 3 and then do 2. In general, the extensive and at times rather intensive exchanges on box 3 have confirmed that this box remains at the heart of our work. While some of you argue this box should be the crux of our efforts and the added value of our three-year work, others are concerned it's creating new law and going beyond our mandate. I believe that there is some truth in both stories. On the one hand, this box contains some of the most essential elements of our work. On the other hand, let me remind you that we are just working on a set of elements and not on an instrument itself, let alone on a legally binding instrument. So I encourage you to look at the text in that light, a set of elements of an instrument without prejudice to its nature and other measures, et cetera. The level of engagement by all of you indicates to me that there is a willingness and, beyond that, an ability also to engage on this text. And I trust that we can and will find a compromise solution, provided you continue to work with each other and on the text at hand. Let me also be clear. There is little time left for discussing great new ideas. we will have to work with what we have now, the product of more than two years of hard work. And while I walk you through the specific changes my team and I have identified, and I see them as possible pathways towards such a compromise, and I think we In the text, if you look carefully, we will have on the one hand, strengthened the language resisting existing IHL obligations, and on the other hand, we have more carefully formulated specific requirements with regard to laws. Now about the paras. 1 to 3. We have made a small linguistic amendment in line with some of the suggestions raised in the room. And I understand that for some this may be perceived as watering down some of the language that we had, while for others the changes might not be far-reaching enough. Again, let me say the following. It seems to me that the use of any weapon system that is not in compliance with IHL, including the laws, is undesirable to every High Contracting Party. And I hope that this is the fundamental baseline we can all agree and build on. Furthermore, this text also does not impose any new obligation under IHL, since we are all bound by the rules laid out in these paras already. We are simply stating them as applicable in the context of laws. At the same time, I would like to remind delegations of our mandate, which is to formulate a set of elements of an instrument, not the actual instrument, as I said before. As such, the formulation we have picked seems to us to be more in line with that task, while retaining the notion of impermissibility of such uses in order to not water down existing obligations. In other words, that explains we moved from "prohibited" to "must not". We have also added a small addition to para 3 to address the concerns by some on the change of status of civilians, and I hope that this may address this concern. Para 4, the retention clause, as it's been now colloquially called, and this power has drawn a rather extensive level of discussion. In the end, it's my belief that this conversation is probably better suited as future work, rather than addressing it here today again. Because of that, I have decided to remove the second sentence and bring the sentence more in line with existing language. the remaining sentence. Para 5, this para has been rephrased as we see this as a key existing requirement under IHL. Now on para 6. This is to most, if not all, present here today the heart of our work, and while I understand that the text that we have discussed in previous versions, and probably also the one in front of you today, may not yet be acceptable to some or acceptable to all, I want to appeal to all of you to constructively engage on this text. Ultimately, What makes laws unique is their ability to function without a human in their critical functions, thus raising a number of serious concerns which are also far-reaching, have far-reaching consequences. It's our collective responsibility to address these concerns and to ensure that the necessary guardrails are established, namely by ensuring the retention of the role of a human. At the same time, I hear the concerns of those delegations that note this does not constitute an existing rule of IHL and feel that it should not become one. Let me therefore recall and let us all agree once and for all that here today we are not creating new law. Rather, we are elaborating elements that may or may not in the future be used to that end. or to an entirely different end, and that the discussion should be approached accordingly. In this context, I would also like to appeal to delegations still having difficulties to accept this text to continue working with me on the basis of this text, and to find a way to address the concerns raised by the greater part of the international community, and to meaningfully reflect the human element in our work. The suggestion made from the floor earlier this week was to use the existing text while offering further clarity on what the concept of context-appropriate human judgment and control is and to what end it should be used, a means to ensure compliance with international law. You will also see that I have added a life cycle approach to this para, which is to make this concept more holistic while clearing up some text in box 4. Para 7. Para 7 is then the operationalization of para 6. It contains the ways in which context-appropriate human judgment and control can be effectuated and through which international law, in particular IHL, and the principles and requirements can be upheld, which is something we try to further clarify with the revised wording. The addition of contextual factors and situation has been introduced as a way to further clarify the term of context appropriate. The sub-paras have then been amended to read more clearly as possible measures. Now about the sub-paras. Subpara remained largely unchanged while responding to certain concerns raised with regard to the reflection of ethical consideration. Subpara was amended in line with some of the proposals heard in the room to bring it more closely in line with language found in existing law, while adding a level of clarity by focusing on limiting the effects. We have also chosen to remove "which includes". While I, like several others, understand the types of targets, duration, scope and number of engagements to be part of the scale of operation, I acknowledge that varying understandings of this exist and that this may be a more accepted wording. Subpara Charlie required some tweaking to bring the text more in line with the technical standards while making it more widely understandable. The change from mission parameters to target parameters does precisely that by using a term that is well understood and limited to the function of concern. The terms real-time and machine learning have been removed to address the concerns by some that this text risks limiting machine learning as such. Picking up on some of the suggestions, we have therefore replaced it with "while active in an operational context" to make clear that the concern is that a loss during an active operation autonomously changes aspects linked to the critical functions without human authorisation. Subpara Delta remains unchanged. Subpara Echo. has been removed as it has become redundant due to the revisions introduced in the previous paras. Its original added value lay in limiting the scope to military objectives by nature, with the intention to capture anti-personnel systems. However, as no agreement could be reached on such a limitation, this wording was subsequently removed. As a result, the para became duplicative since the defined perimeter is now covered by sub-para, while the restriction to military objectives is already contained in IHL and further reflected through para. Para. This has been combined with para in box. For some delegation, the integration of para 8 in its previous location created not only substantive concerns, as they saw it as going beyond existing IHL and possibly impeding technological developments, but also structural concerns, as it was not clear to all to which parts of the text the para referred to. At the same time, I have heard strong views on the centrality of this language. Seeking a middle ground solution, I have suggested to integrate the components of this para into box 2. I have done this with some reluctance, as box 2 seems to me to be part of the text that enjoys the widest range of acceptance. Nonetheless, I believe that this change may be agreeable to you, as the concept is retained while the unclear aspect has been removed. And with this explanation, I would now like to open the floor to hear the views of delegations and their comments. The floor is open. So I'm going to ask for the behind this number. I will now make some of that. We have a first volunteer, Norway. Thank you very much for opening the floor.

Thank you, Chair. I'll do this chronologically, box 2 first and then box 3. So, thank you very much for the suggested reformulation of paragraph 4 with elements of paragraph 8 in box 3. In our opinion, that catch-all is a good solution to the problem of recurring paragraphs with relatively similar purposes. In our understanding, the formulation makes it clear, although maybe only implicitly and depending on the contents of the paragraphs that follow, that those who design, develop and use laws must be mindful to enable as well as to exercise context-appropriate human control of the systems. We miss the word "prohibition" in this paragraph, though, which is important, we think, in light of the content of box 3, which includes prohibitions, and I'll come back to that just now on box 3. We much appreciate the effort that you have put into finding a workable compromise. We note that the wording in paragraphs 1, 2 and 3 has changed from "it is prohibited" to "must not be used". As we see it, the normative content that is expressed in the paragraph remains the same; the difference only lies in the deontic formulation. "must not be used" expresses a positive obligation to abstain from something, whereas "is prohibited" expresses the corresponding negative obligation of prohibition. We could have also formulated the same norm as a lack of permission. From a normative perspective, those are equivalent expressions of the same substantive rule. That said, we still have a strong preference for retaining the phrase "it is prohibited". As we have understood it, only one or two delegations expressed a preference for shifting to the alternative deontic formulation, and we are not quite sure why. Our reason for preferring "it is prohibited" is simply to support the two-tier structure that the GGE has consistently aimed for, with prohibitions, that is, negative obligations, and regulations, which are positive obligations. So, keeping the wording "it is prohibited" preserves that conceptual structure, as opposed to what we have now. In paragraph 4, we note that the second sentence has been deleted, and we thank you for that. We can support that as is. Regarding paragraph 6, we note that it has been changed in subtle but significant ways, we think. As we have stressed many times, the role of human judgement and control sits at the core of the laws problem complex and is essential to how we meaningfully address that. The way that the concept of context-appropriate human judgement and control is framed in our text is not just important; it is essential, and for us it is a particularly sensitive issue. In our view, the current version presents the concept in a markedly softer manner than before. The necessity of human judgment and control for achieving IHL compliance is considerably less explicit in this formulation, and moreover, we are concerned that the formulation "so that the use of laws upholds compliance with international law, in particular IHL", carries an unfortunate connotation, as if IHL was something to be supported or bolstered by the use of weapons. We think it's the other way around. The use of a weapon must comply with IHL, not uphold it. So the correct formulation here, we think, would be use of laws complies with international law. If we are able to resolve the crucial paragraph this week and reach an informal understanding in principle, we could consider going along with the new formulation, albeit reluctantly, because we need to find ways to converge and because we appreciate that, despite the unfortunate connotation that I just mentioned, this is not the spirit of the change and the text. That said, if we cannot do that this week, we would like to see the formulation of 18 December retained for now. Paragraph 7, we note a number of changes. First, regarding the chapeau, we think you've come closer to a formula that could bridge the various positions in the room here. In our opinion, a modal verb would have made this into a clearer sort of regulatory language, but we appreciate the difficulties that that entails. and in light of the intimate connection between paragraphs 6 and 7, we believe that a more descriptive approach, like the one that you have chosen with your wording in paragraph 7, could be supported. We think that the new paragraph 7 does a good job of fleshing out the meaning of the concept that we find in paragraph 6, while at the same time providing a good bridge to the various measures that follow below. It highlights that human judgment and control require contextual awareness. significant concern, though. In the splitting up of the concept of context-appropriate human judgment and control, something very important has been lost, and that is the word "appropriate", which speaks to the quality and the extent of the human judgment and control that are needed, and which, in turn, is determined by a range of operational and other contextual factors across the life cycle. Also, the phrase "contextual factors and situation" seems to suggest, and wrongfully we believe, that the situation is not part of the context, but maybe we're misreading this. By including situation, we're adding unnecessary ambiguity to the text. Now, we assume the point of this formulation is to give the widest possible scope to the context that could be relevant. But in case we believe this part should be reformulated, either by sticking with contextual factors only, or with a different alternative that I will come back to in a moment. Finally, the phrase "is assured through, inter alia, the following measures" could be taken to suggest that the measures that are listed in sub-paragraphs to are either necessary or sufficient, or both. That is, while there might be other relevant measures that can assure context-appropriate human judgement and control, the measures that are listed are sufficient in themselves As we see it, the words "inter alia" do not resolve that issue. I am not sure that that is the intention behind the drafting, but there is real ambiguity. Although we certainly believe that the measures that are listed are necessary, especially in the somewhat optional language that many of them are now couched in, we do not think that it is right to suggest that they are sufficient. The previous formulation underscored that the measures were important but did not allude to any sufficiency, and we do not think that it would be correct to be that categorical. To address the issues that I have just mentioned, we would like to propose a few modifications to the sentence. For simplicity, I will read the entire sentence and then explain it. It would read, and we include the word "appropriate" again, "appropriate human judgment and control depend on contextual factors and then we can add "across the life cycle of the system" if we want to broaden the scope. "and are underpinned by, inter alia, the following essential measures." We believe that, with that sentence, we clarify four things. Firstly, that contextual factors inform the quality and extent of human judgment and control that is needed. Secondly, that contextual factors must be taken into consideration in the proper exercise of human judgment and control. And thirdly, that context must be construed. Not sure where it cut, but anyway, four things that this formulation clarifies. So I think I was at the third when it cut, that context must be construed broadly. And fourthly, that the sub paragraphs list important, but non-exhaustive measures to ensure there is context appropriate human control and judgment. For us, 7A is okay. In 7B, we notice that the measure no longer is to ensure that the effects of laws can be adequately anticipated. With the deletion of that, we lose the important aspects of predictability and reliability. In our opinion, that is fundamental for appropriate human judgment and control and, by extension, to compliance with IHL. We believe that that should be retained as a measure, either in 7B or as a separate subparagraph. Also, in section 7, we think that the word "can" in the part of the sentence that reads that those responsible for the use of laws can limit their effects should be deleted. As it reads now, the measure is to provide the operators with the ability to limit the effects, whereas, if we remove the word "can", then the sentence turns into a recommendation that they not only have such an ability but that they actually impose such limits on the system. I have to say that that would be a very light recommendation. It does not specify how much, just that there should be some limitation. Finally, we regret that 7E is gone. Laws that only target objects raise much fewer issues than ones that target people. a recommendation, and I have to underscore a recommendation, not a requirement, that laws primarily be used to engage objects would be very useful in this context. So we would very much like to see it included again in some shape or form. Thank you.

I thank the distinguished representative of Norway for sharing his views with regard to the text proposed and proposing also amendments. Next on my list is the United States. Let me pass the floor to the distinguished representative of the US. You have the floor.

Thank you, Mr. Chairman. I will only address Box 3, paragraph 6, and the chepo of 7. And we'll come back later on other provisions. Mr. Chairman, my delegation made a number of proposals on box 2. We have heard you, and in the spirit of trying to find consensus and trying to focus on your text here as closely as possible, we have a number of suggestions that we would like you and others to consider this afternoon. To respond to your other point, these are not new ideas. These are ideas at the center of our earlier efforts to find a solution, and we are doing our best to continue to work with everyone to try to find consensus on this difficult issue. We join other delegations that have expressed concern about the term "context appropriate." We appreciate the Chair's efforts to try to address the concerns. However, from our perspective, the revisions do not make the terminology acceptable. For example, human judgment and control are actions, not requirements. The phrase, in our view, does not reflect what is needed to comply with international law. For the U.S. delegation, the key requirements The essential components of IHL compliance are good faith and due diligence. This is outlined in our draft articles proposal, which has been submitted multiple times to this group, including last year in 2025, and joined by a number of other delegations. As reflected in the Vienna Convention on the Law of Treaties, good faith is a general requirement in the performance of obligations under international law. In particular, under IHL, when making any decision governed by the law, commanders and other decision makers must make a good faith assessment of the information that is available at the time. And in 2019, the GGE concluded by consensus that compliance with the IHL requirements and principles requires inter alia that human beings make certain judgments in good faith based on their assessment of the information available. So our first proposal is to replace context appropriate, a term which we have understood to come from the domestic policy of one particular state, and instead replace it with a fundamental element of compliance with international law, good faith. However, Mr. Chairman, good faith is not enough to comply with international law. States, parties to a conflict, individuals must not only have good intentions, they must also follow through with due diligence. As esteemed jurists from the International Law Commission have explained, due diligence is a secondary concept for legal implementation. It is a standard by which to assess the conduct of a state in meeting its primary obligations. What due diligence is required varies according to the circumstances, including in light of the primary rule in question. In our view, this concept addresses an issue that many delegations have raised regarding nominal human input. Cases of IHL violations may very well involve situations in which a person nominally has complete control over the system. They nominally provide some sort of input and can exercise any of the controls, but they are not exercising due diligence in the implementation of legal requirements. They are not applying due care. Mr. Chairman, we support the goal of explaining what is needed to comply with IHL, but we have a principled legal position that context-appropriate human judgment and control is not the correct terminology. Therefore, we propose to replace context appropriate with good faith and control with care. So the sentence would read, good faith, human judgment, and care across the life cycle is needed so that the use of laws. And here, We have a similar perspective to that which was reflected by the distinguished delegate from Norway, that upholds is not quite the right phrase here. So we would prefer to return to something more like is in compliance with international law, in particular IHL, including the principles and requirements of distinction, proportionality, and precautions in attack. The next sentence we would propose to read, "The good faith, human judgment, and care that must be exercised in the circumstances depends, inter alia, on the specific rules and principles that are applicable. To these ends, this may involve, inter alia, the following measures. For us, we say, the good faith, human judgment, and care that must be exercised Because we are grounding this concept in the requirements of existing international law and referring to the specific rules and principles that are applicable. We also propose to refer to the circumstances because that is established terminology from international law. Legal compliance depends, often depends on the circumstances, as our UK colleague explained yesterday. We will come back again with further points on the other paragraphs, but we wanted to propose our general vision for how paragraphs six and seven could be acceptable. And I'll read our proposal out one last time for colleagues. Good faith, human judgment and care across the life cycle is needed. so that the use of laws is in compliance with international law, in particular IHL, including the principles and requirements of distinction, proportionality, and precautions in attack. The good faith, human judgment, and care that must be exercised in the circumstances depends, inter alia, on the specific rules and principles that are applicable. To this end, this may involve, inter alia, the following measures. I thank you, Mr. Chair.

I thank the distinguished representative of the United States for stating the position of his delegation. 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 very much appreciate your hard work and efforts to bring us closer to fulfilling our mandate and formulate a set of elements. On box 3, we would, like Norway, prefer the previous formulation of paras 1, 2, and 3, as this is indeed prohibited under existing IHL. We can, though, in a spirit of compromise, accept these changes if it brings us closer to consensus. The same goes for para. 6. We would prefer the previous formulation, but can accept the changes if it brings us forward. We also welcome the change of para. 5. In para. 7, We can accept the changes made on the measures, but the new chapeau raises some questions for us. We are still studying it and are interested to continue to hear other States' understandings and suggestions of refinements. We believe that it can be refined to be clearer both in relation to para. 6 and to the subsequent measures. We are, for example, not certain that "is assured through" is the best wording here. We are open to different ways of formulating the chapeau and want to emphasize that this is an important part of the text. Finally, we believe that the changes made to para. 4 have gone in the wrong direction. Our concerns relating to accountability, training, research and the development of countermeasures remains. Thus, we cannot accept para 4 as it stands, and we still see that an option would be to delete it as a whole. I thank you, Chair.

I thank the distinguished representative of Sweden for stating the positions of her delegations and for making proposals and also being prepared to accept certain texts, although not considered to be preferential for her. Next on my list is nobody. I see the Russian Federation. Let me pass the floor to the distinguished representative of the Russian Federation, and I see also Turkey. Do we follow? You have the floor.

Thank you, Chairperson. We would like to make a number of comments on section 3, some of them are general and some of them are more specific, on more specific paragraphs. To start with the general comments, we Thank you for the work that has been done on the text. We see this as another step towards achieving consensus and are striving for a golden mean in the context of the proposals and positions put forward by delegations previously. What we would say is that, of course, we are somewhat disappointed by some of the corrections done in this new version of the document, but at the same time, we are quite pleased with a number of the amendments that have been undertaken in parallel. Now, to speak specifically of the paragraphs, we welcome the stylistic amendments in paragraphs 1, 2 and 3. as we see it, they make it possible to give the subparagraphs the character of the relevant mandate of the group, without prejudice to the outcome of the discussion, nor the further steps to be taken in this format. On subparagraph 3. However, we would propose deleting the words "protected as such". We understand that this might be important to a number of delegations, but our delegation has done quite considerable work on seeking any similar wording in IHL that is in force, we've looked carefully at 51, 52 of Protocol I to the Geneva Convention, but finding such wording or similar was not something that we were able to do. That is a serious problem for our delegation in the context of aligning this paragraph 3 with existing international humanitarian law. And given that, we take the view that within this format, we cannot, in the context of paragraph 3, which appeared precisely on the basis of articles 51 and 52 of Additional Protocol I, we cannot try to go beyond the limits of the existing standards of international humanitarian law. And so we propose deleting the words "protected as such". As we see it, if we look at it in substantive terms, this does not add anything to the subparagraph. It just confuses matters. And without it, we keep the same meaning in the subparagraph. It will clearly determine the impossibility of using lethal autonomous weapons systems against the civilian population as such, as well as individual civilians and civilian objects. On paragraph 4, like the delegation of Sweden, we do not feel it is possible to keep the first sentence in this sub-paragraph. Overall, we believe that this paragraph, we felt this before and we continue to feel that this is superfluous, because we have the relevant provisions in the document under 1.4. on a legal review, the results of which, under those results, the relevant weapons systems cannot appear, and so they cannot be manufactured, nor acquired or stockpiled. And therefore, this subparagraph is superfluous, and we Propose deleting it entirely, particularly since we have one, two and three, paragraphs one, two and three, which in substantive terms fully cover this matter, we believe. Furthermore, we are prepared to agree with the amendment with paragraph five and to strengthen the language under this subparagraph. We do not have an issue with clearly confirming within as in the form of a stronger wording the concept of lethal autonomous weapons systems being used only in accordance with IHL. This concept for us is unshakable, it is logical, and in that connection, we have no objection to strengthening the language in this subparagraph. On 6 and the chapeau to 7, and the use of the term "human judgment and control". We have mentioned, we have spoken about this many times and we appreciate the efforts that have been undertaken in the chapter to seven to try and find a compromise solution. We believe that the way that this has been done in seven should serve as a basis. But This does not resolve the issue with the terminology in 6, where we have everything as we had it before and it hasn't been changed. And we would reiterate the use of the term "context appropriate". Just like for the US delegation, for us it is problematic. We would prefer to move away from that wording. On the term "human judgment and control", Overall, we can live with that. And we can be flexible on any other options regarding that concept. We are happy with the term human judgment, which has been used many times in reports of the group for previous years. And we hope that As it was stated by the US delegation yesterday, the commitment in previous reports remains. And as we see it, the term human judgment is entirely acceptable to all delegations in this room.

So we're.

Also prepared to live with this term. Turning to context appropriate and how it is set out in paragraph 7, the logic, the approach as how this is done, does suit us. However, it is our view that we need to bring in some clarifications To be more precise, additional clarification, additional language to make sure that this understanding, this concept is fully set forth, that it be as clear and precise as possible. In particular, it's our view that we need to move away from using the term contextual factors. Contextual factors isn't fully, one we can fully understand in our delegation. We discussed different kinds of contextual factors in the context of this group. And it's our view that we really need to be clear on what we're talking about here. We talked about the situational context. And that's clear to all, we can understand that. Or the operational situation, the environment, the operational environment. That's also clear to us. And in this context, with also some individual delegations have talked about the characteristics of the systems themselves, which provide a context, a specific context. And that element is also one that we can understand, it's quite clear. and all three elements are acceptable, which I've just mentioned, to fully flesh out what is meant by contextual factors or the appropriate context. So we believe that it would be worth still working on the text to seek that clarification. And if that is done, then we consider that from the Russian Federation delegation, there should not be any objection. So we do need to work on the phrase contextual factors and situation in the chapeau, the heading of paragraph 7. To return to paragraph 6, the phrase across the life cycle does not completely suit us. Given the understanding that was achieved previously in this group, we talked about ensuring human judgment at particular stages in the life cycle of these weapons systems, at various stages. This is fixed in the report that we have, 2019. And we believe that that was no accident, that it was established and included in the report because at various stages in the life cycle, the full human judgment is required or not required. So here we're prepared to consider as a compromise, the formulation from the 2019 report, so that would be to change across, to change it to at various stages of the life cycle. Turning to the verb upholds, which is used in this paragraph. In general terms, we feel that the previous formulation was more correct, more appropriate. Which was "ensuring compliance". We feel that's easy to understand. We feel that's precise. But if there is a consensus on the verb "upholds", then we would be prepared to consider this formulation and to seek agreement on it. Then, turning to the subparagraphs of paragraph 7. On, as before, we feel it to be sufficient to just have the first sentence in this subparagraph. The second, we believe, is superfluous, especially as it contains reference to ethical considerations. Our position is consistent here. We are not prepared to agree to understandings in this text and concepts which refer to ethical considerations. And the only reference to ethical considerations, as we already recalled, is contained in the reports of the GGE. We're prepared to follow that, but no more than that. On subpara B, We consider this subparagraph to require further clarity. This subparagraph, we believe, should clearly the limits on the scale of the operation, the type of target, duration and geographical scope of the operation. We do not feel it to be expedient to refer to limiting the impact of the limits, to limit the impact, their effects. It seems to us that we should already talk about limiting specific elements concrete factors and aspects of the functioning of these weapons systems. Also, as before, we do not feel it to be necessary to refer to limiting the number of engagements of these weapons systems. As before, we fail to understand exactly in what way that would help us to achieve better human control. It's our view that it would not do so in any way at all. Turning to subpara C, we do not consider that this reflects the current reality in practical terms. It is our view that it does not reflect reality and it should be completely deleted. subparody. We do not consider it relevant and necessary to refer to deactivation and self-destruct, self-deactivation or self-neutralization mechanisms. These are different ways to deactivate a system, a weapons system, and it's our view that each state can itself identify the best way to have weapon systems deactivated. And also, if we delete this final phrase in D, sub-paragraph D, then it would not in any way restrict those delegations and states that wish to use these mechanisms and the application, they could use them without any problems. So, we would suggest that we delete the end part of sub-paragraph starting with "and/or". Thank you, Chair.

I thank the distinguished representative of the Russian Federation for commenting on the text and making proposals. Next on my list is Austria, to be followed by Turkey and Panama. But let me first pass the floor to the distinguished representative of Austria. You have the floor.

Thank you very much. Thank you for providing us with a new text in order to move forward and find some solutions to some problems. I think some of the amendments that we see have been taken up because of the concerns of some delegations, but I think what we will hear from me and maybe also from other, not all delegations, think that those changes are needed and that the previous text has been in quite a good shape, as many have said in the course of this week. So we will, in some instances, also say that we will rather prefer the previous version, but I will go through the paras one by one. When it comes to 1, 2, 3, I mean, we also understand that the substance is more or less the same. But as Norway, we would also have a preference for a design, let's call it this, that is more similar to other protocols that we are already having. And I think that was always the idea. And especially two and three have been modeled after Protocol 2. And since we are in the CCW context, and since our mandate also has this element that we should take them into account, I think that has been the right way. And as it has also pointed out, it reflects more clearly the two-tier approach which many delegations support. We also have a question mark on protected as such. We don't see that this adds much. We understand the idea, but this idea is more or less already there, I mean, through the deletion of under all circumstances. To go the extra mile to make it even more explicit is a little bit over the top, at least in our point of view. Coming to For this is a part that is important for us, it is one that adds value and, and actually, uh, also adds an element that is maybe not that explicit in, in existing IHL, so it's, uh, it's something that actually makes our work more valuable in general, um, it also is quite clear that you should, uh, not, uh, manufacture, stockpile, or transfer weapon systems that cannot be used in compliance with IHL. Why this is disputed is also a little bit of a question mark to us. And maybe states could explain why they would think this is necessary. The argument that you should pick up and test and other weapons that not exist yet should not exist and only very bad faith actors would actually throw around. I think that's not enough to have it in a text, which in general does not provide too many new prohibitions and regulations. If we would have a much more ambitious text, we could talk about this. But with the current version, we should not even be more restrictive than the existing On five, we welcome the change. I think that's accurate and correct. In six, we also have some question marks. I think the changes that we had in the previous text, where we took the very clear language from box two and added it in box three, was the right one. We are modifying it now a little bit back and are adding elements that we had previously there that were also disputed by other delegations. And this is probably not the right way across the life cycle, actually also creates some question marks, more of a logical way than on substance for us. As others have pointed out, we also have the notion now that the use of laws is needed to uphold compliance of international law, in particular IHL, which is probably not correct. So there needs to be a different formulation. So all in all, para 6 as it was previously was probably the better solution. Seven, we understand the idea of this para to, I mean, the original idea as we always have seen this para is that we take the concept that was laid out in six and specify how it can be achieved. We understand that the idea of this being a concept is not something that some delegations can agree with. So a sentence that makes it more into an active way of what exactly to do is something we can look into. How this exactly can be done, I think is something that we will spend a lot of time this evening. The current formulation does not really work for us because it's missing some elements, as Norway has pointed out, and it is also not putting the following measures in the right light. I mean, something that we could think of, but this is just preliminary. I mean, we will have to think a bit longer about this, especially since there is the problem on how to get the appropriate in, as Norway has pointed out. It could read like human judgment and control will be achieved through inter alia, the following measures taking into account contextual factors and situation. But as I mentioned, the appropriate is still missing. But this could be more direct and it's actually pointing to these measures as something that has to be done so that human judgment and control can be achieved. That human judgment and control requires something is logically also a little bit a complicated concept. We also believe that the following subparas would need maybe a more active elements. I mean, now they are sitting there a little bit like concepts. I mean, the notion seems to be that they are measures, but the ensure that we had in previous version actually made more sense to us as it was actually an active element and something that states have to do. We know that the assured is now in the seven, but as I said, our concept of this whole seven would be more that human judgment and control is something that needs to be achieved through measures that you have to take. And in this case, this would be measures to ensure this and that. On the proposals to add, to change out context appropriate with good faith and due diligence, I think most of us know this from treaty law or as a concept of pacta sunt servanda. which might apply to some extent if we want to apply it to the principles of IHL that are relevant for laws, but probably the right place to do this is rather in box 2. And if you want to lay out these fundamental principles, sorry. Box 2 is rather the proper place. I also want to point out that the original, the origin of these concepts in the draft articles as were presented by a group of states, also there, it's under Article 2 and the title of this Article 2 in the draft articles is cardinal principles of IHL regulating the use of an autonomous weapon system. So if even there they are described as cardinal principles, maybe also in the rolling text, they can be treated as cardinal principles. and not as something that actually governs human judgment and control in an operative concept. I think that would be rather wrong and also a little bit counterintuitive since we are talking about treaty law and very important principles, of course, that also have their origins in Vienna, as we all know. Sorry for being so long, but we, I have a few more points. Um, we are very much attached to the, uh, uh, just coming back to the good faith and due diligence, something that we said in the course of this week already, that it could be of course also, also added as a measure under seven, but as I said, the better place is probably as under box two. Um, on seven A, um, We are very much attached to the ethical considerations. We want to keep it and the solution that you have here at the end of box A works. But at the same time, we also think that this is something that is so fundamental for our work that it also has its place maybe in box two. And maybe this is also the place where it would be more appetizing to other delegation. about this just being part of our reports. I also recalled this week that this is also something that is in the solemn declaration from the 2021 review conference as something to guide our work. So it has a certain standing as a political commitment already and something that should be continuously in our work as it was laid out there. And we are still in the cycle of this review conference that I'm talking about for now. On B, I think there's also some changes that don't really maintain the original idea, because what was done here at the beginning was that we take a principle of IHL and we explain again how this can be implemented. And the idea here, which also comes from the 2023 report, is that you take certain measures so that the human judgment and control to anticipate the effects of a weapon system. So there was also this kind of chronics. can be better achieved through limiting their effects and through taking certain measures to limit the scale of the operation, types of targets, duration, and so on. The idea was not that someone who is responsible can do that. The idea was always that this is an option how human judgment and control can be achieved. So this link or this chain or this logical chain is somehow broken in the current version and I think the previous one. was much better in that regard. On C, we regret that the real-time machine learning has gone, as I always explained. I mean, the formulation that was here has never prevented real-time machine learning. It has been a practical measure that is actually quite practical and that most militaries would probably implement anyway because it's the right way how to use machine learning in an operative context. And I think the problem was just that the word "real time" was not understandable to everyone here in this room. I mean, if possible, we would rather have that back. We also regret that we have lost some important element in E, as Norway has pointed out. If we have the possibility to also look there into something that is more into measures that can be taken to minimize the likelihood of harm to civilians or damage to civilian objects, we would very much appreciate this, because this element is now missing under 7, and it has always been an important part of our work. On 8, for us the right place to have, uh, this idea of design and development and use and that there's, that the prohibitions and requirements of this, um, text have to be taken into account, the right place is box three. For us, this is a crucial element and one of the added values that overall this text has, that it is very clear that also in the design and development, you have to make sure that IHL is upheld, can be complied with. I mean, across the life cycle is not-- is having the same notion, but it's missing a little bit the concreteness. It's much clearer if we talk about concrete stages in which this has to be taken into account. So we prefer the formulation that was original here. And just to say that also in the draft articles, already in article one, in one of the first points, article 1.1, it says autonomous weapon systems must not be designed to And also, Article 1.2 in the draft article says autonomous weapons systems may only be developed such that. So all the other delegations have confirmed that these are key measures in upholding IHL. So they should not be just cardinal principles or general principles under box 2. They need to be very much operationalized as part of box 3. Sorry for being so long, and thank you for your attention.

I thank the distinguished representative of Austria for stating the position of his delegation, commenting on the proposals and also on the views put forward by other delegations. Next on my list is Turkey, to be followed by Panama, Ireland and the Republic of Korea. Let me first pass the floor to the distinguished representative of Turkey. Madam, you have the floor.

Thank you, Chair. As was the case in the previous iteration of this box, we have very similar concerns to Sweden. Since we will discuss paragraph 8 in the context of box 2, I will focus on paragraphs 4 and 7. Having said that, our primary concerns about the content of paragraph 8 remain, notwithstanding its new location. Now on paragraph four, we reiterate our call for changing the word shall to should. Shall carries clear mandatory legal force equivalent to a hard prohibition or obligation. Paragraph four deals with manufacturing, acquiring, stockpiling and transferring laws, activities that are deeply embedded in national defense procurement cycles, alliance obligations and export control regimes. Imposing a mandatory shall refrain obligation at this stage, before we even agreed on a final characterization of laws in box one, would be premature and legally overreaching. The scope and definitions of the possible measures we're discussing remain unsettled. Therefore, binding states to shall language before those foundations are agreed risks creating unintended legal commitments. "Should" preserves the normative weight of the provision as a strong commitment while reflecting the reality that this text is part of an evolving framework, not yet an agreed text. Many CCW instruments have used graduated language precisely to allow progressive development of obligations over time. Regarding the proposal for deletion of the second sentence of paragraph 4. This phrase is not a loophole. It is a deliberate carve-out that reflects a practical military and legal reality. States must understand restricted weapons in order to defend against them and train personnel to recognize them. Without this language, one of two dangerous outcomes follows. States either secretly retain non-compliant laws with no governing framework, creating dangerous ambiguity, or the restriction is applied so absolutely that states cannot develop countermeasures, leaving them vulnerable to adversaries who ignore the framework entirely. The restriction and the exception must coexist. The restrictions prohibit the manufacture, acquisition, stockpiling, or transfer of non-compliant laws. The exception permits retention strictly for training and countermeasures in minimum quantity and temporarily. Without that exception, the restriction appears unworkable in practice, undermining state buy-in, inviting outright rejection of the framework, and losing credibility by ignoring operational reality. A restriction with no exception forces noncompliance. A bounded, accountable exception encourages genuine adherence. Paragraphs 6 and 7 of box 3 require human oversight across the entire life cycle. Training personnel to recognize, counter, and deactivate noncompliant systems is directly consistent with this. Retaining a minimum number of such systems for that purpose allows militaries to test countermeasures against real systems enables operator training in identification and deactivation consistent with paragraph subparagraph 7D and supports the chain of command requirements under paragraph subparagraph 7A. Without this phrase, that training capacity disappears, undermining the human control architecture Box 3 is built around. Alternatively, we can park this discussion for later and delete the paragraph in its entirety for the time being. Now moving to chapeau of paragraph 7. Here, we would like to replace the words "is assured" in the chapeau with "can be reinforced through." "Is assured" implies a deterministic guarantee that human judgment and control will definitely produce a certain outcome. But this is both technically and legally problematic. No state can guarantee outcomes in complex operational environments, and embedding such language creates an impossible standard of compliance. It also shifts the framing from process obligations which states can commit to, toward outcome guarantees which they cannot. The measures listed in subparagraphs a through e are procedural safeguards, not certainty mechanisms. The language should reflect that. Last but not least, sub-paragraph 7. I feel the need to repeat that limiting or restricting the number of engagements that laws can undertake autonomously will create ambiguity and open the door for different interpretations. Allow me to explain in detail what we mean by that. Turkey supports meaningful human control over laws and does not seek to undermine the core objectives of Box 3. Rather, our position is that the phrase, "the number of engagements that laws can undertake autonomously" is technically ambiguous, operationally unworkable, and legally redundant, and that its deletion strengthens rather than weakens the instrument. Paragraph 7 already requires that those responsible for laws used must be able to limit or otherwise restrict the types of the targets, the duration of operation, and the geographical scope. These three parameters taken together functionally constrain the number of engagements without explicitly naming it. If a loss is limited by target type, duration, and geography, the number of engagements is implicitly and necessarily bound. The term number of engagements is not defined anywhere in the rolling text. This creates serious problems. Does one engagement mean one target identified, one target selected, one target fired upon, one target neutralized? In a fast-moving, dynamic battlefield environment, the line between a single engagement and multiple engagements may be impossible to determine in real time. Does a failed engagement where a target is selected but not struck count toward the number? This definitional vacuum means that compliance with this requirement would be impossible to assess, either by operating state itself or by any external verification mechanism. We warn that vague obligations are dangerous obligations. They invite selective interpretation and undermine the uniform application of the future framework across the board. Paradoxically, specifying number of engagements as an explicit control parameter could weaken human oversight rather than strengthen it. Here is why. If a state pre-authorizes a loss to conduct, for example, 50 engagements, a human operator may feel absolved of further responsibility once the number is set, treating the numerical limit as a substitute for ongoing situational awareness. This creates a false sense of compliance. A state could claim it exercised human control by setting a number, while in reality exercising no meaningful judgment about individual targeting decisions. True human control as contemplated by paragraph six requires contextual and continuous judgment, not a one time numerical authorization that then runs automatically. Deleting this phrase in sub paragraph 7B redirects to focus towards the qualitative contextual human judgment that paragraph six demands, rather than a quantitative checkbox that could be gamed. Thank you.

I thank the distinguished representative of Turkey for stating the views and concerns of her delegation and making proposals. Next on my list is the distinguished representative of Panama. Panama, you have the floor.

Thank you, Chairperson. I'll be brief on paragraphs 1, 2 and 3. We agree with other delegations that have spoken already, namely that we need to go back to the original language. On 4, we have reservations regarding the change carried out in the first sentence, but we would support deleting the second for the reasons that we already stated during previous exchanges. On paragraph 5, We welcome the change made to "must", which we believe strengthens the message based on the principles of international humanitarian law. And finally, in 6, 7 and 8, we would also prefer to return to the original language. Thank you.

I thank the distinguished representative of Panama for commenting on the proposals made and reflecting on it. Next on my list is Ireland, to be followed by the Republic of Korea. But first, the distinguished representative of Ireland, you have the floor.

Thank you very much, Mr. Chair. We're very much still digesting these changes, but we're happy to provide some preliminary remarks as well as some reactions to some suggestions made from other delegations. As a general remark, we also prefer the previous version of box three. In paragraphs one through three, we would prefer the prohibition formulation, as said by Austria and Norway, including other delegations, as these are in line with other protocols under the CCW And more importantly, it's the core of the two-tier approach of prohibitions and restrictions. We're also concerned that these changes, taken in combination with the changes in box 2, paragraph 4, which says that below are requirements and measures, could risk weakening existing IHL. So we strongly support returning to prohibition language. Regarding paragraph 4, we welcome the removal of the retention clause for the reasons expressed by ourselves and many delegations. We still maintain that "refrain" in the first sentence is too weak. Any systems which are incapable of being used in compliance with IHL should not exist, and therefore a clear prohibition is necessary here. So we would suggest for the first sentence to read, "It is prohibited to manufacture, otherwise acquire, stockpile, or transfer loss, which are incapable of being used in compliance with IHL," and the rest would read as written. We also welcome the change in paragraph five from may to must. In paragraph six, we appreciate the efforts in finding a landing zone here. We can support Norway and the United States and others who suggested to move away from the word uphold and instead say that the use of laws complies with or is in compliance with international law. And we also welcome the inclusion of the across the life cycle. Like Austria, for us, we prefer using the terminology streamlined across the long text of design, develop, and use, which we think is more specific and offers more clarity than lifecycle. So with that together, with some, a couple of other tweaks, the paragraph could read, context appropriate human judgment and control, including in the design, development, and use of laws, is required so that the use of laws complies with international law, in particular IHL, and the rest reads as written. We appreciate the proposals made by the US here. We do believe they were made in good faith, judgment, and care. An immediate reaction is that we cannot live without a reference to human control in this paragraph, which is obviously the crux here for many. For us, due to the nature of autonomous weapon systems, in order for a human to exert human judgment and care as well as due diligence, some level of human control is needed. Paragraph 7, we understand this is an effort, again, to accommodate some concerns by some delegations, but we do think that with the changes, this paragraph, which is intended to serve as a bridge underpinning the two-tiered approach, could be compromised. We have some similar concerns with Norway here regarding losing the important term appropriate. And we also recognize some tension introduced with is assured. So we can positively look at Norway's suggestions here. Perhaps another option instead of is assured through, we could explore, for example, must be operationalized through, inter alia, or must be exerted through, inter alia, or something to that effect. 7B, we can also-- we also take issue with the word "can." We think it's important to maintain at least some limits here. So we suggest its deletion. And previously, the concepts of predictability and reliability were present to some degree in terms of the words "adequately anticipated and controlled," which are now gone. So we again strongly urge the reinsertion of predictability and reliability in this section as these are foundational for ensuring compliance with IHL. 7C we think is headed in the right direction. We do suggest to reinsert mission so that it would read target and mission parameters. And we also wish to retain paragraph E. We do not see the difficulty with restricting the use of laws to military objects. and should certainly also maintain the minimization of likelihood of harm to civilians or damage to civilian objects. We had earlier suggested language on avoiding civilian harm, but at least we want to retain what is in paragraph E, and we support maintaining paragraph 8 in box 3 for the reasons highlighted by Austria. Thank you, Mr. Chair.

I thank the distinguished representative of Ireland for stating the position of his delegation and making comments and suggestions. Next on my list is the Republic of Korea, to be followed by Switzerland, Australia, Colombia and Israel. But first, the distinguished representative of the Republic of Korea, you have the floor, sir.

Thank you, Chair, and we appreciate your efforts to provide the new suggestion that may allow us to move forward. I will focus on my intervention on paragraph 4, 6, 7, presenting my delegation's preliminary views. First, paragraph 4. We share the concerns expressed by Sweden, Russia, and Turkey regarding the paragraph 4. Our position on the need for the reference to the exceptional cases for the training and development of countermeasures remain. We believe that the need for this retention clause is eloquently explained by our Turkish colleague. Furthermore, for the same reason that you, Chair, replaced the word "it is prohibited" with "must not be used" in paragraph 1 through 3, we suggest replacing "shall" with "should" in paragraph 4. At this stage, it may be helpful to organize the elements under the categories of "must" and "should" and without introducing the additional categories. However, also as some delegations and you, Chair, have suggested, if this issue should be addressed at a later stage, we would propose deleting the paragraph in its entirety for the time being, as suggested by our Swedish and Turkish colleagues. Furthermore, we can also consider the suggestion made by our Norwegian colleague yesterday to retain only the reference to manufacturing and stockpiling, as those should remain the focus of our discussion if we do not wish to address the issue of retention for the purpose of developing countermeasures. Paragraph 6, we expect a more in-depth discussion tonight on this concept of context-appropriate human judgment and control. So at this stage, I will simply reiterate our delegation's reservation against on this term. We also note the suggestion by the US delegation to consider alternative formulations such as good faith, human judgment, and care. We appreciate the continued constructive engagement of the United States in helping us to identify language that could eventually facilitate consensus and enable us to produce a set of elements as we are mandated. In our view, the approach the Chair introduced in the chapeau of paragraph 7 may be more workable, namely unpacking the notion of context appropriate rather than creating a new and potentially ambiguous term. So moving on to paragraph 7. With regard to the chapeau of the paragraph 7, we believe that the phrase taking into account the contextual factors and situation could be replaced with the language we already agreed by our group in guiding principle C, namely the operational context and the characteristics and capabilities of the weapon system as a whole. So we don't need to reinvent the wheel. Regarding paragraph 7B, we can view the paragraph as amended favorably in the spirit of compromise, despite the fact that we still have a reservation regarding the term number of engagement. Our reservation on this term, again, is well explained by our Turkish colleague, so we don't need to repeat. However, we would like to point out that we do not support further changes to the paragraph, structure of the paragraph. We understand the measures listed after restricting as a non-examples, and therefore consider it appropriate to retain the verb "can." We also welcome the revisions to paragraph 7C. particularly the deletion of the reference to real-time machine learning. That said, the paragraph remains somewhat awkward in its current formulation, may require some further streamlining. Lastly, we welcome the merging of the paragraph seven of box three, sorry, paragraph eight of box three with the paragraph four of box two, which improves coherence and logical structure of the text. Thank you, Chair.

I thank the distinguished representative of the Republic of Korea for commenting on the text and sharing views and making proposals. Next on my list is Switzerland, to be followed by Australia. But let me first pass the floor to the distinguished representative of Switzerland. You have the floor, sir.

Thank you, Mr. Chair. We support all efforts aimed at bringing us closer while preserving the core elements, and we also appreciate the work you have invested in doing so. While we understand the need to remain creative, we hope to focus our efforts on building on the existing text rather than introducing entirely new formulations at this stage. Following the two-tier approach, setting out prohibitions and regulations, we regret the changes in paragraphs 1 to 3 for the following reasons. One, this change could be interpreted to mark departure from working towards a legally binding instrument, ideally a CCW protocol. Second, for many, it is important to keep existing IHL language and adapt it to lethal autonomous weapons systems. We share that view, while we also see the need to develop the law where law's specificity so requires. The language in paragraphs 1 to 3 now diverges from the standard IHL language, especially also looking at CCW and its protocols. Certainly, we also see a risk that such change weakens existing and longstanding formulations of prohibitions on which these paragraphs are based upon. We would therefore prefer to keep the original formulations. That said, we fully understand what you said about the role of this rolling text, that its finality remains open and that we should address the text with that in mind. With this understanding, we could, at this moment, live with adaptations from prohibitions to must not. Regarding paragraph 3, we have important concerns about the changes made as they introduce a new formulation changing long-standing and widely accepted IHL language of existing IHL rules, such as Article 3, paragraph 7 of Amended Protocol II of the CCW, or Article 2 of CCW Protocol III. These formulations have been acceptable in the past, and for us, it's not clear what special aspects of laws would require changing it now. This also considering that many delegations expressed their wish to remain close to existing IHL language. If not mistaken, the formulation proposed is not used in other IHL treaties. it therefore risks to include a new formulation that may be prone to different interpretations. We do understand concerns expressed in regard of this paragraph. However, we do not see a fundamental change with laws that would require this addition, for instance, compared to anti-personal mines. Of course, we could include language that specifies, for instance, when civilians lose their protections from attacks, or define what exactly is a military objective. But this should be based on existing language. Moreover, doing so would raise the question why we do not restate other rules and principles, for instance, that in case of doubt, civilian status must be presumed. For these reasons, we would revert to the previous versions of paragraph 3. On paragraph 5, We welcome the change from may to must, reflecting the strong and convincing calls made in the room. International humanitarian law does not present compliance as optional. On paragraph 6, we would like to reiterate that for us the concept of context-appropriate human judgment and control forms the basis of our work and is key. Like others already said, there are some linguistic issues around uphold, but we are sure this can be solved easily. For instance, with a formulation like context appropriate human judgment and control across the life cycle of laws is needed to uphold compliance with international law and international humanitarian law, or it is needed so that the use of laws is in compliance. On paragraph 7, year 2, we think we are getting closer. Just make sure that we create clarity what exactly needs to be assured. In addition, the formulation and in quote and situation is not clear. So, while we are open to further study how this could be solved and to further reflect on an explanation, what exactly context appropriate human judgment and control encompasses, we believe this is not yet ripe and potentially could be treated separately in the future, for instance, also in a definitional article. Overall, we note that box three, which lies at the core of these elements, should retain sufficient ambition and granularity regarding the measures to be taken and their underlying rationale. This is important for us to maintain that laws and their effects can be adequately anticipated or the restrictions to clear perimeters with a view to protect civilians and civilian objects, even if implied in the geographical scope. We will come back on paragraph 8 when we discuss box 2. Chair, in light of the limited time remaining, we hope that all delegations will make every effort to seek compromise and convergence as we approach the end of our mandate. The time ahead will require flexibility, focus, and a shared willingness to move closer together. We thank you for taking the text forward. We note that many elements have been improved and that others have been streamlined or removed where they did not attract sufficient agreement. We appreciate your efforts to reflect the discussions in the room and to keep the process moving forward in a constructive spirit. Thank you, Mr. Chair.

I thank the distinguished representative of Switzerland for sharing the views of his delegation, commenting on proposals made by other delegations and for his words of encouragement to the Chair. Next on my list is Australia, to be followed by Colombia and India. But first, the distinguished representative of Australia. You have the floor, Madam.

Thank you, Chair. Thank you also for sharing textual proposals for further discussion today. We welcome a number of these edits, including some proposed by my delegation. While we reserve our position, we appreciate the opportunity to offer some preliminary remarks and engage in an interactive discussion on your proposals. First, Australia welcomes your inclusion of "protected as such" at paragraph 3. In our view, the words "protected as such" resolve an inconsistency with IHL. The paragraph now states clearly that civilians may not be targeted when they are protected as civilians under IHL, but does not extend protection to civilians that directly participate in hostilities. Turning to paragraph 4, we appreciate your explanation that you have removed the second sentence on the basis that it is better addressed in future work. As this text only represents elements of an instrument, we expect that we can go along with your proposal in the interests of consensus. Chair, on paragraph six, Australia greatly appreciates the changes made. In our view, this better represents how, across the life cycle, context-appropriate human judgment and control could assist states and personnel to uphold compliance with IHL. As a grammatical point, we see the inclusion of "across the life cycle" as best located after "is needed" so that it reads "context-appropriate human judgment and control is needed across the life cycle so that". Chair, in respect of paragraph 7, we share the concern of Norway that the amendments have lost the link between human judgment and control and what is intended by "context-appropriate". As we have explained, in our view, The appropriate measures of human judgment and control to be applied will depend on what is needed in the context to support compliance with IHL. The earlier text of paragraph 7 better captured this intent, and the link between paragraph 6 and paragraph 7 was clearer. We also share concerns related to "assured". As Norway explained, this could be seen as suggesting these measures are sufficient. We are also concerned that this would be understood as meaning these measures are required in all circumstances, which we understand is not the intent. We would therefore prefer that we revert to the prior version of paragraph 7, but may suggest alternatives as our thinking develops. We also express our thanks to other delegations for offering suggestions today, which we will definitely reflect on. Chair, we are still studying the edits you have made to the subparagraphs in 7 and reserve our position on those for now. but we support the deletion of paragraph 8 and its incorporation into box 2, paragraph 4. We will have some comments on how it has been incorporated into that paragraph 4 when we discuss box 2. I thank you.

I thank the distinguished representative of Australia for stating the position of her delegation and giving preliminary comments on the text. Next on my list is Colombia, to be followed by India and Israel. But let me first pass the floor to the distinguished representative of Colombia. Madam, you have the floor.

Thank you, Chairperson, and thank you for your efforts and those of the secretariat. This is a difficult but necessary exercise, and we're very close to the end of the mandate of the group. We have some comments on this box. First, like many other delegations, We would prefer in 1, 2 and 3 returning to the original language, namely it is prohibited in all circumstances for a number of reasons. Firstly, this is clear and direct, in line with what IHL provides for, and is in line with the language of the protocols of the Convention. It is prohibited. That is a clear prohibition which gives no scope for interpretation. Second, we welcome the elimination of the second sentence in paragraph 4. That was something that we discussed in previous comments. Third, in paragraph 5, we agree with the change from may to must, and we could also accept shall, although we believe that many delegations would not be able to support that. So we can accept the change to must. 4. In reference to paragraph 6 and the word "uphold", many delegations have commented on this, and we would agree with the delegations who would prefer a more direct, legally clear wording, namely context appropriate is needed so that the use of laws complies. And likewise, for legal clarity, we'd like to have a reference to design, development and use instead of life cycle. As a fifth point, and this is in 7A, we refer to the wording that you have proposed. This gives clarity on the consideration of ethical considerations and on 8, that will be discussed in box 2, I believe, but for our delegation, it's important to have a general reference to the design and development of these weapons systems. That will be in box 2, and I am sure that we will discuss this more when we come to box 2. But as has been said, this is important. It's important to assess this provision For our delegation, we need to have a general provision like this in the rolling text. It needs to stay somewhere. And we did not comment on the previous box in its revised version. So we would just like briefly to reaffirm that we agree with many delegations which made comments this morning on the second sentence in 1A. It is important to have this, and we would need to have the proposed amendments included there. Thank you.

Let me thank the distinguished representative of Colombia for stating the position of her delegation and making proposals and reacting to the views of other delegations. Next on my list is India, to be followed by Israel, Cuba, Egypt and Spain. Let me first pass the floor to the distinguished representative of India. So you have the floor.

Thank you, Mr. Chairperson. We wanted to just take the floor to appreciate the good work that you've done in other sections, but particularly in box three. And we thank you for all the edits that you've introduced in this box. And we believe that changes that you've suggested are both reasonable and practicable and that they certainly take us in the right direction. And we also acknowledge that achieving political consensus as well as technical and legal clarity is a very difficult bar to achieve. But what we want and how we look at it is that we want a good enough set of measures that seek to address the most important concerns which are raised by laws and adopt corresponding measures that can help us in ensuring compliance with international humanitarian law. We have a few observations on box 3. These are not very substantive points, but I wish to go through them one by one. On para 1 to 3, we thank you for attempting to calibrate the language with the IHL language. We also noted that there were comments by some delegations on some further gaps that remain. So we would request you to further edit if that is possible. On para 4, these changes are also acceptable to my delegation. However, I think that this paragraph could benefit from some more clarity that research and development, which is not in the nature of manufacturing, stockpiling, and acquisition or transfer, would not be impacted. We had also suggested in our last intervention that the paragraph 8 covers the same operative meaning but has a more palatable formulation, which of course you've removed now. But we had suggested that paragraph four and paragraph eight are linked in some ways, and eight has a more palatable formulation which could be acceptable to all other delegations. On para six and para seven, our view is that para six and seven are, that these two paragraphs must be integrated in some way. because they talk about this notion and this concept about human judgment and control. And what I observe is that both these paragraphs have different formulations, which might, you know, be misinterpreted to create two different benchmarks. We appreciate, however, your changes in para 7. This is a suggestion that we had proposed, and we thank you for introducing that in the text. We did hear comments from certain delegations around ambiguity, which contextual factors and situations might introduce. It is right that this term is not defined in the IHL. But we do believe that IHL principles implementation and its cardinal principles by themselves are context dependent without an explicit statement or explicit expression that is given in the IHL. And like some delegation pointed out, we do believe that this context could have at least three, encompass three different things, including operational context, circumstances at the time of use, and the weapon characteristics. And these, and given that there is complexity around what these contextual factors are, and also given the fact that there is a lack of a better canonical term, we could also, for instance, use the word contextual considerations if that makes it easier for some delegations to accept this motion. And with that, the suggestion on paragraph 7 would look as follows. Human judgment and control should take into account contextual considerations through inter alia following measures. And we are, of course, open to any other terms that are suggested by the delegations to capture the complexity of this term without really introducing that directly or explicitly in the reference. Now, on para 7, our view is that, like the chapeau paragraph 7, para 7 should also be tethered to the notion of human judgment and control and it being taking into account contextual considerations. because this is a view that we had expressed last time when this box was discussed, that direct human intervention will not be possible in many occasions, including for reasons such as time sensitivity. For instance, when an autonomous weapon system accidentally moves into a civilian concentrated area, it should be able to change its target parameters to limit unintended engagements and limit damage. So human judgment and control, in our view, seems better, but we will be happy to discuss further, and of course, we understand that there's going to be discussion on this aspect today in the evening, so we look forward to that discussion. And my delegation also appreciates your preserving reference to ethical considerations. I see that you have, in order to accommodate the concerns of some delegations, you have removed the word "due regard" and used "taking into account", so we and also relevant ethical considerations, so my delegation can certainly support that reference. And last but not the least, we also support perhaps a separate paragraph referring to concepts in IHL which are implicitly derived, such as on good faith and due diligence, while we admit that these terms are not used directly as such, but we believe that the IHL sets a certain standard, which includes performing obligations in good faith, the requirement to verify targets, and taking feasible precautions, including under Article 57. And together, these provisions establish a standard of conduct requiring commanders and states to actively assess the information available at the time, exercise reasonable judgment, and continuously ensure compliance with IHL during planning and conduct of military operations. And we are also willing to use relevant language as is contained in IHL to elaborate such a concept if there is movement in that regard. Before I end, we just want to thank you for the text that you've given. My delegation is flexible in most aspects in box 3, and we feel that we are moving in the right direction, possibly closer to an agreement in box 3. Thank you.

I thank the distinguished representative of India for commenting on the text and for elaborating proposals. Next on my list is Egypt to be, oh no, sorry, Israel, Cuba, Egypt, Spain, and Brazil. But let me first pass the floor to the distinguished representative of Israel. You have the floor.

Thank you, Mr. Chair, and good afternoon, colleagues. We want to thank you, Chair, to you and your team for all the efforts in providing us with additional suggestions for box 3, and we will share our initial remarks regarding them. Regarding paragraph 4, although we recognize the constructive revision, as other delegations have already noted, in our view, this is still not inappropriate paragraph for the text as we think it should maintain its focus on the use of laws in order not to stray from the CCW's objective and expertise as an IHL-focused treaty. We will therefore request its deletion. Regarding paragraph 6, we would like to be very clear regarding the concept of context-appropriate human judgment and control. This term remains a serious concern for us. Israel has consistently opposed the inclusion of this term since it was first introduced in the discussions and throughout the meetings. As we have listened carefully over the past few days to the statements made by other delegations, the obvious lack of shared understanding of this meaning, on the meaning, sorry, of this concept only reinforces our concern. We believe that the text should be revised in a way that better reflects language capable of commanding broader support and consensus among delegations. Unlike distinction, proportionality, and precaution in ATAC, which are established operative rules of IHL, context-appropriate human judgment and control is not itself a rule of international law. Introducing such ambiguous concepts risk creating legal uncertainty rather than facilitating the implementation of existing law. From an operational perspective, the concept is equally problematic. The level and form of human involvement in the life cycle of laws may vary significantly depending on the system. the operational environment and the mission, while still fully complying with IHL, provided that the applicable rules, including distinction, proportionality and precaution in attack, are respected. Although we recognize that some delegations may view contexts appropriate as reflecting this flexibility, in our view, It risks papering over divergent understandings of the term, which later on will lead to conflicting interpretations and ultimately undermine, instead of bolster, commitment to IHL. For the purpose of moving the discussion forward and getting closer to a consensus, we are willing to positively consider the proposal put forward by the United States, which in our view represents a constructive compromise. Alternatively, in the spirit of constructive engagement and finding a workable solution, we will suggest the formulation that we mentioned in our previous intervention, meaning measures, including during development, relating to control and human judgment are needed to ensure the use of laws is in compliance with the principles and requirements of distinction, proportionality, and precaution in attack. Moreover, as previously noted, we believe this issue will be more appropriately addressed in box 4 rather than in box 3, which should reflect primary rules of IHL. Regarding paragraph 7, in our view, the change introduced in the chapeau moves us further away from consensus and does not reflect the concerns raised in previous discussions. Firstly, the use of the word "requires" risks suggesting that these measures constitute a mandatory legal condition for lawful use, which is not supported by IHL. Framing these measures as requirements could therefore unintentionally imply that the failure to implement one of the list of measures will render the use of the system unlawful, even where the attack itself fully complies with IHL. This will introduce a new legal standard that does not exist in the law. Second, the paragraph is assured through is particularly problematic as it implies that the listed measures are the mechanism through which lawful conduct is guaranteed. In reality, these measures may support compliance, but they cannot guarantee it. Compliance ultimately depends on the overall design, operational parameters, and the manner in which the system is used in a specific attack in accordance with IHL. Depending on the circumstances, compliance may be achieved even without applying the listed subparagraphs or by applying only some of them. Third, we reiterate our concern regarding the repeated reference to human judgment and control. For these reasons, we support the proposal made by the United States for paragraph 7, which provides a more balanced and legally sound formulation. Alternatively, if this paragraph remains following paragraph 6, we suggest shortening the chapeau to read: "Such measures can include, where appropriate, the following." Moreover, as we previously mentioned, this paragraph lists a variety of practical measures, some of which may be needed to implement IHL in certain circumstances and some that reflect policy measures that exceed legal obligations. Considering the practical nature of this paragraph, the appropriate place for it is box 4, which outlines measures that support compliance with IHL rules. Regarding paragraph 7, we believe that the words "for example" should be added after "including" or, alternatively, items from the list should be deleted or qualified significantly, as the examples given are only practical tools that may be used in specific circumstances to limit the effects of loss, rather than a mandatory requirement in every situation. Furthermore, there may be other measures that can be applied, and this is merely a list of potential examples. Finally, regarding paragraph 7c, We consider this paragraph particularly problematic and not in line with how we believe the lawful use of laws should be implemented and addressed. First, it is not sufficiently clear to what the term "target parameters" refers to in this context, nor what would constitute a substantial modification to such parameters. As currently drafted, the provision would therefore be difficult to apply in practice. Second, the phrase "are authorized by a human" when combined with the wording "while active" may be interpreted as requiring continuous or real-time human intervention. As we have stated previously, such requirement does not reflect a rule of international law. That's all for us for the moment. Thank you.

Let me thank the distinguished representative of Israel for stating the provisional positions of his delegation. Next on my list is Egypt, to be followed by Cuba, Egypt, Spain, Brazil, Japan, Italy. But first, the distinguished representative of Cuba. Madam, you have the floor.

Thank you, Chairperson. We thank you for your efforts and those of your team on the text. We recognize the complexity of this exercise. We're now analyzing the proposals, but in preliminary terms, we have some comments to make, like Norway, Sweden, Panama, Colombia, and other delegations that have already spoken. We would prefer the previous version of paragraph 123 in the wording "it is prohibited", which better corresponds to a two-level approach that we have been working with up to now, and which my delegation believes gives greater clarity. On paragraph 4, we would agree with the proposal presented by Ireland for the start of the paragraph, namely it is prohibited to manufacture, acquire, stockpile, or transfer laws and so on. Furthermore, we would also agree with the proposal to eliminate the second sentence in the paragraph, as in the version that you have presented. And we also support the proposal to change paragraph 5 as displayed. We cannot support removing a human judgment, context appropriate human judgment and control in line with the comments made by the US delegation. And in paragraph 7, the ethical considerations remain relevant here. We are concerned by the changes made in paragraph 7 chapeau and various sub-paragraphs. Like the delegation of Norway and others, we would prefer to have a more regulatory approach, as we had in the previous version. We sincerely hope that this evening's debate brings us closer to a solution on paragraphs 6 and 7. I thank you.

I thank the distinguished representative of Cuba for stating the views of her delegation, and I can only support your last remark. Next on my list is Egypt, to be followed by Spain. But let me first pass the floor to the distinguished representative of Egypt. You have the floor.

Thank you, Chair. My delegation would like to express its appreciation for your continued efforts guiding our work. In reference to paragraph 6, delegations expressed differing views regarding the legal status and placement of human judgment and control concept. Some emphasized its central importance for ensuring compliance with IHL, while others noted that it should be understood primarily as a mechanism through which existing legal obligations are implemented. Chair, while we understand the intention behind the context appropriate human judgment and control concept, we still believe that the inclusion of context appropriate may introduce unnecessary ambiguity and subjective interpretation, as many delegations indicated in its interventions. That's why it's important to discuss further this issue, especially taking into consideration that the term the group will reach will be used as literature. in other ongoing and future fora dealing with laws. And while Egypt reiterates that, uh, uh, the crucial, it's crucial to guarantee, uh, meaningful human judgment and control, our delegation earlier on, uh, on Tuesday proposed re- removing the ca- uh, qualifier context appropriate before human judgment and control. We therefore propose framing the provision so that it explicitly states that human judgment and control over the critical functions of identification, selection, and engagement are required to ensure compliance with IHL, while clarifying that the nature and degree of such control should be appropriate to the operational context. Therefore, we suggest para 6 reads as follows: "To ensure compliance with IHL, in particular, the principles and requirements of distinction, proportionality, and precautions in attack, human judgment and control throughout the life cycle of laws is required. The nature and degree of such judgment and control must take into account the type of system, its anticipated effects, and the operational environment. On the new amendments made on first three paragraphs by removing the prohibited term, We join other delegations who expressed their desire to retain this language as it goes along with the two-tiered approach that calls for prohibitions and regulations. I thank you, Chair.

I thank the distinguished representative of Egypt for stating the positions of his delegation and commenting on the text. Next is Spain, to be followed by Brazil and Japan. Let me first pass the floor to the distinguished representative of Spain. You have the floor.

Thank you, Chairperson. Thank you for your work and that of your team. Regarding section 3, we have the following comments to make. In paragraphs 1, 2 and 3, we believe it is appropriate to retained the previous text, namely, it is prohibited in line with the language used in other protocols to the CCW. This also reaffirms the two-tier approach of prohibitions and regulations. We understand that there is a set of factors here, but we could accept the new tenses under the previous premise. On 4, refraining from manufacturing and acquiring, we are attentive to the doubts put forward by Sweden, but we believe that it is important to have further debate on this. On paragraph 6, The first thing is that human judgment is fundamental for my delegation. In order to implement the principles of responsibility and accountability, there needs to be sufficient human judgment and control in the use of these weapons systems in order to assess the correctness of the use. The second issue is the inclusion of the -- across the life cycle we believe is appropriate. The human machine interaction takes place over the life cycle and can be done in different ways. This does not mean that it needs to take place at every stage. In paragraph 6 and 7, on the operational context, we're more favourable to human judgment and control, as we set out in paragraph 6. According to the proposal from the Republic of Korea for paragraph 7, this is agreed language in the guiding principles. It would be acceptable for my delegation, therefore. It would be appropriate to keep the same language in paragraphs 6 and 7. This, what is relevant, is covering operational context. The context in which these systems are used is fundamental and thus guiding principle six on human machine interaction, which depends on factors such as operational context or characteristics and capacities of the lethal autonomous weapons systems, is essential. And as we have reiterated on many occasions regarding B, we have a reservation on the limiting this is not covered by any conventional or customary norm. On 7C, we welcome the change to significantly modified. We believe that it is much clearer now. Thank you.

I thank the distinguished representative of Spain for stating the positions of her delegation and commenting on the text. Next on my list is Brazil, to be followed by Japan and Italy. But first, distinguished representative of Brazil. You have the floor, sir.

Thank you. Thank you very much, Mr. Chair. I do appreciate your efforts and your team's efforts in offering these alternatives. Like others, we have a strong preference for the previous formulation for paragraphs 1, 2, 3. We believe that the language here should be aligned with other international humanitarian law instruments, and in particular with the CCW protocols. We are also somewhat concerned with the new element that was included in paragraph 3, namely the reference to protected as such after civilian objects. As others have noted, this is not in line with formulations that have been used in other IHL instruments. On paragraph 4, we welcome the changes made, although we do note that there are remaining concerns, valid concerns, with the lack of a specific exception. We are open to considering any alternatives, but we do believe that the first part of this sentence, of this paragraph should remain. On paragraph five, we welcome the change introduced. Turning now to paragraph six, like other delegations such as Austria, Pakistan, Ireland, Colombia, and others, we would prefer a clearer reference to the issue of human control and judgment as a necessary element for IHL compliance. On paragraph 7, like Australia, we believe that the link between this paragraph and the preceding one became somewhat less clear. We understand and welcome the objective of the changes that were introduced in the second part of this chapeau to make a reference to contextual factors and we believe that this is an important element, but there might be ways to increase the clarity of this chapeau. We are considering alternatives and we intend to offer them during the consultations tonight. On the subparagraphs of paragraph 7, we believe that the changes were mostly appropriate and contribute to enhancing the clarity and the conciseness. But we note the concerns expressed by other delegations, in particular Ireland, with the loss of the reference to anticipation of effects. And we hope that this can be addressed either by tweaks to this language here or on box four. And I thank you, Mr. Chair.

I thank the distinguished representative of Brazil for his comments on the text presented, suggestions presented. Next on my list is Japan, to be followed by Italy and Canada. But let me first pass the floor to the distinguished representative of Japan. You have the floor.

Thank you, Mr. Chair. I would like to share our preliminary comments on paragraphs 6 and 7. On the term "context-appropriate human judgment and control", We listened carefully to remarks of delegations and we recognize that there are various understandings and interpretations. We believe that clarifying common understanding of this concept is important, not only in fulfilling the mandate of the GGE, but also for the future work of this body, and that there is merit for all in finding workable solutions. Also, as stated by other delegations, we should build upon the consensual work of the GGE. In this regard, we appreciate the constructive engagement and are open and willing to consider all proposals, including those from the United States, and look forward to further discussions tonight. On context appropriate, we consider that unpacking this term would be a good way forward towards consensus. Contextual factors and situations in paragraph seven seemed somewhat synonymous or vague. And we support the proposal by the distinguished delegate of the Republic of Korea to use agreed language from guiding principle C and spell out factors to enhance clarity. Lastly, on 7B, we still do not quite understand how we count the number of engagements in practical terms. and concur with Turkey and the Republic of Korea and others on its deletion. I thank you.

I thank the distinguished representative of Japan for commenting on the text, reacting to proposals made by other delegations. And let me now pass the floor to the distinguished representative of Italy, to be followed by Canada. But first, Italy, you have the floor.

Thank you, Chair, for your hard work on this section, for which we express our general support. Allow me, however, to offer a few comments and react to some of the proposals made by others. On paragraphs 1 to 3, we previously supported the wording "it is prohibited", as we consider it more consistent with the prohibitions and regulations that this section seeks to establish. Nevertheless, I will listen carefully to the explanation provided by other delegations. We can accept the formulation "must not be used" in the spirit of compromise. Regarding paragraph 3, we are flexible on the phrase "protected as such." We believe the primary aim of IHL is to protect civilians, and the current language reinforces this objective. However, removing that wording would not substantially alter the meaning of the sentence. On paragraph 4, we note that the concerns raised in our previous intervention have been addressed, and we can support the current language. On paragraph 6, we appreciate your efforts to bridge divergences. We remain flexible regarding across the life cycle, should delegations prefer to specify individual phases. We would nevertheless prefer the use of "comply" in place of "uphold compliance". We have taken note of other language proposals as well, and we wish to reiterate our readiness to work constructively with delegations to reach a landing zone. At the same time, we wish to emphasize the utmost importance of the human element. We deem essential that it continues to be reflected in this paragraph. On paragraph 7, We would have preferred to retain the word "anticipated", but in a spirit of flexibility, we can support the current version. Regarding paragraph 7, we support your proposal, though we consider that "self-destruct", "self-deactivation" and "self-neutralization" are already encompassed by the expression "deactivated" in a timely manner. We will revert to paragraph 8 when it is discussed under section 2. And I thank you, Chair.

I thank the distinguished representative of Italy for the comments made. Next on my list is Canada, to be followed by Utrecht University and the United States. But first, the distinguished representative of Canada. You have the floor, sir.

Thank you, Chair. Thank you to you and your team for the efforts to bring the many views that we have in the room with these proposed edits to the rolling text. In box three, we appreciate the reworking on the paragraphs here and can largely support what has been proposed. Please allow me to share my delegation's views on some of the paragraphs. Firstly, on paragraph three, we agree with Australia and others and the inclusion of protected as such. On paragraph four, we believe the deletions are acceptable. Weapon systems that cannot comply with international humanitarian law would be inherently unpredictable and unreliable and therefore unlawful to use. As such, there would be no need to consider them further in the context of countermeasures and development. Moving to paragraph 6, we would first like to repeat Canada's position from earlier this week. We do not see context appropriate human judgment and control as an IHL obligation. We see it as a necessary measure to ensure compliance with IHL principles. Elaborating on paragraph 6, My delegation is generally supportive of the paragraph, and we note the addition of across the life cycle, which my delegation finds acceptable as it helps contextualize when context appropriate human judgment and control is applicable. My delegation sees reference to context appropriate and control as important elements for facilitating IHL compliance and ensuring that armed conflict remains a fundamentally human enterprise underpinned by human accountability and responsibility. My delegation wishes to share some additional acuity in this regard. In my delegation's view, context appropriate is necessary as it allows human judgment to be applied at different times for different reasons throughout the life cycle of laws. In our view, humans must apply their judgment in good faith and adapt their judgment to the context as per current military practice. On human control, my delegation views this as a necessary measure to ensure legal compliance, preventing algorithmic errors, guarding against technical malfunctions, upholding ethical standards and maintaining appropriate responsibility and accountability. With that said, we wish to respond to the proposed new formulation to this paragraph to include good faith, human judgment and care as an alternative to context appropriate human judgment and control. We have taken the time to consider this proposal and We agree that good faith and due diligence are general principles of international law, but interpretation of IHL is not limited to them. Therefore, as a concession to the use of the concept of good faith, we propose the following permutation as an option for consideration. Context appropriate human judgment and control applied in good faith across the lifecycle. Finally, in paragraph 7B, which introduces restrictive measures to limit the use of laws, my delegation questions the precision of the language, which refers to those responsible for the use of laws. We feel that, in line with other language found in the rolling text, this could link to the requirements for a responsible chain of command and control. As such, we offer the alternative formulation. that those in the responsible chain of command and control of laws can, when required, limit their effects, including by restricting the scale of the operations, followed by the remainder of the paragraph as drafted. Mr. Chair, of course, we will share these remarks in writing with you and your team. I thank you.

Let me thank the distinguished representative of Canada for his remarks, observations, and also trying to build a bridge between proposals that have been tabled by delegations. Next on my list is the Utrecht University, and then we have the United States, France, Pakistan, ICRC, Mexico, Stop Killer Robots, Philippines, Algeria, UK. So I think Some of them will have to deliver them during the informal or later tomorrow. We will see how that works out. But let me first pass the floor to Utrecht University. You have the floor, Maaike.

Thank you, Chair. I'm delivering this brief intervention on behalf of the delegation of the University of Southern Denmark and my own, Utrecht University. Chair, with respect to discussions around box 3, paragraph 6, we note that some delegations have made remarks regarding the desire to revert to design, development and use here. However, we believe this does not reflect a comprehensive understanding of the lifecycle of these systems referenced in other places in the rolling text and fails to encapsulate all relevant lifecycle stages important for context appropriate human judgment and control the heart of this paragraph. point, which was just echoed moments ago by the delegate from Canada. From a technical as well as a legal perspective, and following our earlier interventions this week, we again respectfully suggest that retaining language across the lifecycle more explicitly encompasses stages of planning, designing, developing, using, and reviewing. This language echoes an accepted scientific AI lifecycle approach. The language also tracks largely across the legal language in Article 36 of AP1 addressed to high contracting parties in the study, development, acquisition, or adoption of a new weapon, means, or method of warfare. I thank you, Chair.

Let me thank the distinguished representative of Utah University. And next on my list is the United States. Let me pass the floor to the distinguished representative of the US, and then later we have France and Pakistan, but first US.

Thank you, Mr. Chair. Regarding the revisions to the "it is prohibited" formulations in paragraphs 1 to 3, we align ourselves with the view expressed by Italy. Regarding paragraph 1, the phrases "which are" should be "that is." So we should say "a law that is of a nature," and we can send you those edits again. Regarding paragraph two, we reiterate our serious legal concerns. We will not make new proposals, but we can accept a laws must not be used if it is designed to target civilians. Regarding paragraph three, we've explained previously our concerns that have not been addressed. At this stage of our efforts trying to find consensus, we propose to add with the intention after must not be used. We also propose to delete the civilian population as such as well as individual. By referring simply to civilians protected as such, the language would be understood as including both the civilian population and individual civilians. This was the solution to this issue in the context of the Arms Trade Treaty. So the paragraph would read, "Laws must not be used with the intention to make civilians protected as such and civilian objects the object of attack." Regarding paragraph 4, we support the proposals to delete this paragraph. Regarding paragraph 5, we reiterate our previous legal concerns to be as surgical as possible in our edit. We propose to add "applicable" before IHL in paragraph 5. Regarding subparagraph A of paragraph 7, we cannot accept the paragraph as one, but we can accept it if divided into three subparagraphs, because we see these as three distinct measures rather than one measure. So subparagraph A would be operating laws under responsible chain of command and control. A bis would be legal review, including of the effects and of the identification, selection, and engagement functions. A TER would read, taking legal requirements into account when applying relevant ethical considerations. The application of ethical considerations can reinforce and support the implementation of legal requirements. We hope this proposal could address the concern that we're elevating ethical considerations into a new standard separate from IHL and help the group reach consensus on this issue. Regarding subparagraph B, we can support the proposal to delete the first part of the paragraph and also to include the language that the distinguished delegate from Austria read out from the draft articles proposal sent by a number of states. I won't read it out again, but I thank Andreas for reading it out. We would include in addition eight, pardon me, we would include in addition a B bis adjusting the scale of the operation, the types of targets, duration, timing, geographic scope. In a spirit of flexibility and being persuaded by the detailed and thoughtful intervention from Turkey, we can support deleting the number of engagements as we, based on their intervention, can understand this idea as being included in the language scale of the operation. If possible, if it's possible to In terms of consensus, we propose to include timing because adjusting the timing of the operation, for example, to nighttime when civilians are not present out and about is often a critical feasible precaution to protect civilians. We propose to delete subparagraph C. Regarding subparagraph D, we propose to separate it into two sub paragraphs or two paragraphs. We should address in the first one, we should address deactivation and redirection. And we would prefer to be more ambitious than this, but in the spirit of compromise and flexibility, we suggest deactivating or redirecting, which would be subparagraph D. And then paragraph D bis would be incorporating self-destruct, self-deactivation, or self-neutralization mechanisms. For us, it's important to separate these into two paragraphs because they're different measures and they're taken at different times in the life cycle. And lastly, we support the general observation that these measures need to relate more explicitly to the protection of civilians, which is one of the key objectives of international humanitarian law. I thank you, Mr. Chair.

Let me thank the distinguished representative of the US for making comments and suggestions. Next on my list is France, to be followed by Pakistan and the ICRC. But first, the distinguished representative of France.

Madame.

Thank you, Chairperson. Generally, we welcome this new box three and clearly see the compromise solutions that you have identified. We believe that this has achieved a delicate balance. Not everything is perfect, but at this stage in our discussions, we need to be asking ourselves if we can live with certain things. And what we wish to do is to explain our preliminary reading of these paragraphs and state why they could suit us. We continue to see 1, 2 and 3 as prohibitions, but we can be flexible on the way that that is expressed in the text. In 4, we have the same concerns as other delegations. The explanations from the various delegations have allowed us better to understand this. It is important to avoid manufacturing, acquiring, and acquiring laws that cannot be used in compliance with IHL that comes from the prohibition on the use of this specific system, namely entirely unpredictable systems whose effect cannot be anticipated. And retaining this type of system, this very specific type of system, for training or development of countermeasures also seems not very relevant. However, we have heard divergent interpretations, and we could discuss this. Paragraph 7. This allows us to explain that human Control depends on the context and factors. And it is implemented through various measures, which also depend on the context. In A, IHL requires the chain of command to assess the situation and decide on the use and the parameters, depending on legal and operational considerations. but it also decides on a case-by-case basis to consider ethical considerations. In, IHL requires that the tone of command anticipate the effects and limit them. It can limit them, can limit on a case-by-case basis. The parameters of use the type of target, the duration, geographical scope and the number of engagements, depending on its appreciation. And c) the mission parameters must be determined in advance by the chain of command and substantial modifications that can change the context must be reassessed by the chain of command. So, I thank you therefore for this part three, which generally suits us.

I thank the distinguished representative of France for commenting on the text proposals put forward. Next on my list is Pakistan, to be followed by the ICRC. First, let me pass the floor to the distinguished representative of Pakistan. Sir, you have the floor.

Thank you, Chair. We join others in appreciating you and your team. In the interest of time, we would only address a few issues on which we have heard some comments, and we can always come back on the other changes. Chair, on Paragraph 123, we join others in calling for reverting to the previous language. We think that the terms it is prohibited is the standard language. In fact, when we look at Additional Protocol I, different provisions, let's say Article 35 to Article 51, it is also an example that we use in CCW protocols or mandates specifies that. And how we look at it, we see that the prohibition, when we talk about it is prohibited, we are talking about attaching this to the weapon itself. It applies absolutely regardless of who uses it. An obligation that is framed only as must not be used perhaps shifts the framing more from the weapon to the user. And also, Most importantly, we have not yet heard from delegations that favour the change what specific concern they had from the phrase "it is prohibited", for which we have to use the new framing. If the concern is that prohibitory language prejudges the nature of the instrument, and our mandate is to develop elements of an instrument without prejudging its nature, we must note that the argument cuts both ways. If it is prohibited, is said to prejudge the instrument as a prohibition-based treaty, then must not be used equally prejudges it as a measure-based document. Therefore, I think it is not a choice about the nature of the instrument, but it is a reflection of how we use these practices in our CCW and different parts of the IHL. Chair, we also join others in paragraph 3 to express our concern on the phrase "protected as such". We think that instead of adding back the language of "at all times", we have further diluted this paragraph. Chair, on paragraph 4, we respectfully disagree with the language, with the argument that has been put forward by those who call for its deletion. that this language is superfluous by legal reviews. We think that these two things, legal reviews and this paragraph 4, they are two fundamentally different things. Legal reviews are a process. They are the mechanism by which a state determines whether a weapon can comply with IHL or not. And this language in paragraph 4 is a prohibition. It's a consequence that follows after that determination is made. Chair, on paragraph 6 and 7, we think that they form an integrated structure. One cannot exist without the other. Paragraph 6 states what is essential, and paragraph 7 elaborates how it is to be assured. On paragraph 6, we can see positively the change of introducing across the life cycle. We think that this paragraph 6 should be saying it out clearly that context-appropriate human judgment and control are essential for ensuring compliance with IHL. We think the group has reached this agreement before. on human judgment and control separately, and now it's time that we state it together. And we certainly welcome the changes that have been introduced by Norway and supported by others in this paragraph. Chair, paragraph 7, I think it helps what you have tried, unpacking the context appropriate what comes under context appropriate, and we can take guidance, of course, as others have said, from the Guiding Principle C of 2019. We have always believed that the context includes everything. It can hold the contextual factors, the situation, the systems and capabilities of the weapon system. But however, if it needs to be spelled out further, we can spell it out based on Guiding Principle C. But here, I think the word "situation" adds ambiguity, and we can support Norway in calling for its deletion. Chair, on paragraph 6 and 7, a proposal has been made to replace these notions with good faith, human judgment, and care. We agree that good faith and due diligence are important concepts in IHL, but they cannot replace human judgment and control. I think they operate at different levels. Good faith is a standard of mind. It refers to a subjective state of the decision maker. Due diligence is a standard of conduct, whether reasonable steps were taken. Both are important, yes, but they presuppose something more fundamental. And that fundamental is that a human is in a position, is in control to exercise that good faith and that due diligence in the first place. Moreover, good faith applies to every weapon under IHL. It is not laws specific. And we here in this language are trying to identify a laws specific formulation with a general concept that already applies to all weapons. Chair, on the issue of limiting engagements, this would be the last issue on which I comment. before ceding the floor. We have heard some concerns on limiting the number of engagement, that it is ambiguous and it is unclear, and we would like to comment on it. First, we think that this notion is not a standalone or new concept. It is definitely, it is directly borrowed from the precautionary measure that is specified under Article 57 of Additional Protocol I. It requires those who plan or decide upon an attack to take all feasible precautions to avoid or minimize incidental civilian harm. Therefore, we think that limiting the number of autonomous engagements before requiring renewed human input is one practical way of giving effect to these obligations. It creates a checkpoint. Second, we would like to highlight that the text that you have proposed does not prescribe a specific number. It does not say that laws may not undertake no more than, let's say, five engagements or six engagements. What it says is that limiting the numbers of engagement is one of the measures through which human judgment and control is assured. And the specific numbers, of course, will depend on the system, operational environment, target set, and the risk to civilians. Third, I think on the practical concern that different weapons work in different ways, we agree. A defensive system protecting a naval vessel would work in different ways in response to incoming missiles. But this is precisely why the Chapeau frames these measures in a non-exhaustive, in a context-appropriate way. And that's why I think the language that you have proposed, you have retained on limiting number of engagement works. I would end my intervention, Chair, at this stage, but I would just like to flag that we have some additional number of comments on some other language proposals that appear in sub-paragraphs of paragraph 7. Thank you.

Let me thank the distinguished representative of Pakistan for further sharing his thoughts with regard to the text. I think we still have eight requests for the floor, so we will not be able to deal with that this afternoon. So we will have to continue tomorrow morning still with box number three. This evening, we have an informal session for high contracting parties and the ICRC. It starts at 7 in the A building in room 11. Be aware, as I said on Thursday, on Tuesday, that the gate to the UN premises closes at 7:00. So if you go out and you come back after 7:00, you will not be able to enter. Be aware of that. And please note also that delegations who are not in possession of a magnetic badge, they are advised not to leave any hour after 6:00. because you will not be able to reenter the grounds of the UN here in Geneva. And I think that's it from my side. It's clear that, We still, I won't say, have 128 aircraft that have to find a landing spot. Some of you have joined aircraft, but I think there's still a sizable number of different aircraft that all have to land on a pretty small spot. So there's quite some work to be done. For that reason, we have also a discussion this evening on on box 3, especially on the paras 6 and 7, and see if we can narrow down divergences there. I think that would greatly help us in the further discussion of box number 3. Not that there are no further issues, but I think paras 6 and 7 are the ones where we have really to spend some extra time on. Having said that, I wish you a good break and see you back at 7:00 in Room 11 in the A Building. Thank you very much. The meeting stands adjourned.