Group of Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous Weapons System Geneva, 2-6 March and 31 August - 4 September 2026
Provisional agenda 1. Opening of the meeting. 2. Adoption of the agenda. 3. Confirmation of the rules of procedure. 4. Organization of the work of the Group of Governmental Experts. 5. Further consider and formulate, by consensus, a set of elements of an instrument, without prejudging its nature, and other possible measures to address emerging technologies in the area of lethal autonomous weapon systems, taking into account the example of existing Protocols within the Convention, proposals presented by High Contracting Parties and other options related to the normative and operational framework on emerging technologies in the area of lethal autonomous weapon systems, building upon the recommendations and conclusions of the Group, and bringing in expertise on legal, military, and technological aspects. 6. Other matters. 7. Adoption of the report. 8. Closing of the meeting.
Machine-readable formats: Plain text · JSON
Transcripts available through this tool are created by using automatic speech recognition and are not official records nor official documents of the United Nations. Official records and official documents are available on the Official Document System of the United Nations. Learn more
Excellencies, distinguished delegates, thank you for being in the room. I hope you had a good rest. Let me start by thanking the Pakistani presidents of the CD because today is Tuesday and that would have meant a regular meeting of the CD and they have made our life a lot easier by not calling for any meetings this week. So again, I would like to appreciate that very much, also on behalf of you. We continue with the list of speakers from yesterday afternoon when we were discussing paragraph 25, the characterization of laws. And I have one, two, three, four, five, six, seven. I have Germany, Israel, Japan, Norway, Canada, and Chile. Please indicate, and I see now also Belgium, indicate if you would like to take the floor with regards to the characterization because then we have an indication on with regard how many more requests for the floor there will be on this topic before we move on to discussing the human element. I see The list growing now. Good. Having said that, let's now start again with first Germany to be followed by Israel. But let me now pass the floor to the distinguished representative of Germany. You have the floor.
Thank you very much, Mr. Chairman, and thank you very much for launching a very positive dynamic through the additional text proposal you presented yesterday. certainly addressed many of delegations' comments from earlier on Monday and triggered a lively and interesting interactive debate among delegations. I hope that there will be more of that on other parts of the report today. As mentioned, in the spirit of reciprocity, we will try to show utmost flexibility, and I would be grateful if you could put your proposal on the screen. Thank you. Specifically on the text that you proposed, When we go to paragraph 25, the chapeau, operate together might be a simpler way to formulate functionally integrated. We believe that the proposal that was suggested by the United Kingdom to keep the word can in the sentence might be more accurate, especially in combination with paragraph 25b. I think one of the variations could be technological components that, operating together, can identify, select and engage a target, and so on. In our view, the term lethal is already included in the term labelling the object of that characterisation, as we speak of a lethal autonomous weapon system. Therefore, we see a number of delegations voicing their need to include it in the text to balance paragraph 25. In our view, the addition of "with lethal force" after "identify, select and engage a target" in paragraph 25, chapeau, could work, as long as paragraph 25b remains as is and the characterisation as a whole remains balanced. Regarding the moment of engagement in paragraph 25a, we are still looking at this in more detail. We are reflecting on the comments of the colleagues again from the United Kingdom and others on this, and we agree that circumstances might be a better fit. Thank you very much.
Let me thank the distinguished Ambassador of Germany for sharing the views of his delegation and commenting on proposals made. Next on my list is the distinguished representative of Israel to be followed by Japan and Norway, but let me first pass the floor to Israel. You have the floor.
Thank you so much, Chair, and good morning, colleagues. Mr. Chair, regarding the first paragraph of 25, we welcome the changes reflected in the new text. We welcome the addition of with lethal force, which appropriately reflects the focus on the characterization on a lethal autonomous weapon system. Regarding the formulation, technological components that operate together to identify, select, and engage, We recognize that it addresses several of our previous concerns regarding the ambiguity and potentially broad scope of functionally integrated. We are still considering the precise formulation, but we see this change as a positive step. We will also prefer to retain the formulation to identify, select, and engage rather than reverting to can. In our view, two more clearly links Okay. Okay. Sorry, I will repeat the sentence. We will also prefer to retain the formulation to identify, select and engage rather than reverting to can. In our view, two, more clearly links the technological components to the functions that characterise the system. Regarding paragraph 25, however, we are concerned by the addition of and the moment of engagement. In our view, this introduces a new cumulative element to the exclusion. The essential question here should be who determines the target to be engaged rather than the precise moment at which that determination is implemented. Where the personnel planning deciding upon or carrying out the attack has determined the specific target or specific targets, the critical targeting judgment has already been made by the human. For this reason, we believe it is important to delete the reference to the moment of engagement and retain the previous formulation. At the same time, we are considering other formulations proposed by the UK and the US regarding the timing of attack or circumstances of engagement. However, as an initial remark, we believe that circumstances might require a clarifying qualifier to avoid an overly broad interpretation. So these are the comments for this morning. Thank you so much.
I thank the distinguished representative of Israel for commenting on the text and reacting to proposals made by other delegations. Next on my list is Japan, to be followed by Norway and Canada. But let me first pass the floor to the distinguished representative of Japan. You have the floor, madam.
Thank you, Mr. Chair. We would like to thank you and other delegations for providing concrete text proposals to steer our collective work forward. On the addition of with lethal force in the chapeau, we concur with the eloquent explanation by the distinguished delegate of the UK and others yesterday that the revised text does not limit the scope to weapons systems that result in loss of life. It maintains a technology neutral approach in capabilities, And we can support restoring can to further emphasize this point. Read in conjunction with subparagraph Bravo, paragraph 25 as a whole clearly ensures that weapon systems that do not result in loss of life are indeed included. Therefore, we support this revision. Moving to subparagraph Alpha, our delegation views moment of engagement to be on the one hand very specific, which may not be technically feasible to determine precisely. On the other hand, it could be very ambiguous in a sense that it puts into question what engagement means and covers. While we agree that circumstances are relevant and should be taken into account and are open to considering the UK proposal, we would like to recall that we do expand on various factors to take into account as well as measures to ensure compliance with IHL later in the text, notably in paragraph 37 and 38B, as previously noted by the distinguished delegates of Singapore and India. For these two points, we support and prefer the proposal put forward by the distinguished delegate of the United States to refer to the timing of attack, which addresses a temporal issue in a concise and understandable manner and it also fits well with the verb determine in the text. We support other amendments. I thank you, Mr. Chair.
I thank the distinguished representative of Japan for commenting on the text and proposals. Next on my list is Norway to be followed by Canada and Chile. But let me first pass the floor to the distinguished representative of Norway.
Chair, thank you very much. Thank you for these additional suggestions of yesterday. First, we support the replacement of functionally integrated with your formulation as amended by the UK. So that's the weapon and technological components that can operate together to identify, select, et cetera. And I think they made the point very well that, I mean, we need to have some kind of the word can in there in order to not, you know, exclude 25B. from really coming into effect. Second, human being is okay for us. Third, 25a, after having listened carefully to your explanations yesterday and after some extensive talks in the break, we understand that the second sentence excludes from the scope of laws only those systems where the operator has determined one single highly individualized target, so for instance, enemy vehicle with license plate this and that. for example, or where the operator has determined one single highly individualized group of targets, which would be then a collection of targets that have a clear common identity because they are co-located in space and time, but where the individual members of that group have not been individually singled out. For instance, something like that convoy of enemy tanks moving down that road. On the other hand, if the operator has determined more than one target, even highly individualized ones, for instance, enemy combatants A, B, and C, who are located in different parts of the battlefield, were no longer in the second sentence, but in the first sentence of 25A, the trait of being an enemy combatant or an enemy vehicle is not a specific group of targets in this understanding. So from that perspective, and having pondered your suggestions a bit more, we believe your proposal to write individual group of targets is acceptable. However, like Brazil raised yesterday, this exclusion should not come into effect just because the operator has determined a specific target for the exclusion to apply. The specific determination should also have been made with respect to the more precise conditions of the engagement. I think the UK again explained well why the precise sort of moment of attack might not always be so easy to determine in advance. So we would therefore support a reference to the circumstances of the engagement more broadly as a replacement for the wording that you have proposed on the screen there. And finally, as regards the words with lethal force, that is not something that we are in favor of to put more sort of emphasis on lethality, but we could support the Swiss proposals to simply write with force. Lethal, of course, is in our mandate in that it's found in the very concept of laws yet, As we see it, the mandate does not contain a definition of laws, and it doesn't obligate us to give a specific meaning to those words. It's up to us, to the GGE, to suggest to the CCW how this concept should be understood, and we believe that lethality is well and truly covered in the term laws itself in combination with subparagraph b. But if with force could alleviate the concerns of some, then we, so the Swiss proposal, then we could accept that. Thank you.
I thank the distinguished representative of Norway for commenting on the text and the proposals made. Next on my list is Canada, to be followed by Chile and Belgium. But let me first pass the floor to the distinguished representative of Canada. You have the floor.
Morning, colleagues. Chair, my delegation appreciates the effort you and your team have undertaken in helping us find consensus on paragraph 25. We've been listening attentively and would like to react to some of the suggestions that we have heard from the room. On the chapeau, we would like to support the UK's proposal, which has been supported by many delegations, to reintroduce CAN ahead of identify, select, and engage. Additionally, We would support having just force instead of lethal force, which in our view offers better clarity on the types of system intended to be included within the scope of the characterization and is more in line with IHL. Finally, we would like to express support to the UK suggestion in subparagraph a to have circumstances of engage as this would help address the temporal considerations. We can also accept the US suggestion of timing of attack here. Thank you.
I thank the distinguished representative of Canada for sharing the views of his delegation and commenting on proposals made by other delegations. I do not see the distinguished representative of Chile in the room, so we leave that for the moment. Let me now pass the floor to the distinguished representative of Belgium. You have the floor.
Thank you, Mr. Chairperson. My delegation wishes to offer its condolences to the people of Norway for the loss of their head of state. As this is the first time my delegation takes the floor, let me start by thanking you, Mr. Chairperson, and your team for your relentless efforts to lead our work to a successful conclusion. Belgium is generally supportive of the rolling text and we are confident that a consensus will be reached on the text by the end of the week. Regarding paragraph 25 and your new proposal, Belgium shares the concerns expressed by some delegations yesterday that the insertion of the words with lethal force adds a criteria to the characterization and therefore risks excluding certain weapons, namely those that do not inherently possess a lethal or kinetic capability, such as, for example, cyber weapons or the less than lethal weapons. I would also like to recall that adding lethal to the characterization under now para 25 was already a major concession for my delegation. We would therefore prefer to remove this addition from the chapeau, but could also support as a compromise the Swiss proposal made yesterday and relayed by the US to retain only the words with lethal, sorry, with force in the chapeau. Even though any weapon inherently implies a use of force, we could live with this wording if it helps reach a consensus. We agree with your proposal to replace the terms functionally integrated by operate together, but we also believe there is merit in retaining the word can. And in this respect, we support the counter proposals made yesterday by the UK and others. Finally, regarding the quantifier in the second sentence of para 25A, we believe it is important to pay attention to any addition here. It seems to us that adding wording such as and the moment of engagement or alternatives heard yesterday risk excluding autonomous weapons from the qualification from the moment that a human predetermines its specific target or group of targets and the circumstances of its engagements or the timing of the attack. That would, in our view, not be what we are seeking to do here. Thank you.
I thank the distinguished representative of Belgium for stating positions of his delegation and commenting on proposals made. Next on my list is Singapore to be followed by South Africa, but let me first pass the floor to the distinguished representative of Singapore. You have the floor, sir.
Thank you, Chair. Good morning, colleagues. Chair, thank you for giving my delegation the floor again. We return to the topic of paragraph 25 alpha on the issue of moment of engagement, having listened to discussions yesterday. Chair, you had asked for reactions to proposals. We note the proposals made by the UK and the US on the circumstances of the engagement and the timing of attack respectively. We see the merits of both proposals, but have some suggestions with the intention of improving feasibility. Regarding the term circumstances of engagement, My delegation finds it difficult to square away the definite nature of the word "determine" with the issue of circumstances. In the current formulation, the operative word "determine" seems to connotate the ability to control or influence something directly. This will not be possible, given the dynamic nature of the circumstances involved in the execution of an attack. One can set mission perimeters to ensure that the weapon system operates within certain circumstances, but many other situational factors when the engagement actually occurs lie beyond an operator's control. These could range from sudden changes in weather conditions to the actions taken by an adversary. The point is that a state can never really determine the exact circumstances in which a weapon system engages its target, only the intended or expected circumstances. In this regard, one possible formulation of the paragraph could be, and I quote, "A system is not included in this characterization when the personnel planning, deciding upon, or carrying out the attack determines the specific target or specific group of targets. and take into account the anticipated circumstances of the attack. I repeat the addition, the last part of the sentence, it will be and take into account the anticipated circumstances of the attack. Thank you, Chair.
I thank the distinguished representative of Singapore for commenting on the proposals made by other delegations and proposing some new text. Next on my list is the distinguished representative of South Africa. You have the floor.
Regarding paragraph 25, South Africa can support the revised formulation in the chapeau related to the replacement of functionally integrated with the systems that operate together can. In our view, this wording provides useful clarification and appropriately reflects situations where functions may be undertaken in conjunction with or as a part of a broader set of systems and capabilities. It also accommodates considerations relating to the ability of systems to function in an interoperable manner while maintaining the necessary flexibility in the text. We therefore consider this to be a constructive formulation and can support its retention. Regarding paragraph 25a, South Africa appreciates the efforts to bridge different views on this issue. We were supportive of Brazil's initial formulation, particularly the reference to circumstances in which engagement takes place. At the same time, we can also live with the United Kingdom's reformulation referring to the circumstances of engagement as we believe it preserves the essential intent of the provision while offering language that may be more broadly acceptable. Thank you, Mr Chair.
I thank the distinguished representative of South Africa for commenting on the text making proposals. Next on my list is the distinguished representative of Denmark. You have the floor.
Thank you, Chair, for allowing my delegation to take the floor. We thank you for circulating the latest version of para 25 based on the discussions we had yesterday. and we would like to make a few comments in this regard. As regards the CHAPEAU, we support your addition of operating together instead of functionally integrated as we find this formulation to be more precise and easier to understand. Likewise, we have no objection to using the formulation human being instead of human. On the matter of adding with lethal force, we have some concerns. It is the view of my delegation that this term may create ambiguity in the text. By including it, we potentially risk excluding certain autonomous weapons systems from the characterization that ought not to be excluded. We reiterate that rules of IHL governing the means and methods of warfare does not include a criteria of lethality. That being said, we will not stand in the way of reaching consensus, and we are prepared to accept this proposal. In consideration of the fact that the use of a loss that does not result in loss of life is expressly addressed in subpara b, and as we consider that IHL will be applicable regardless of this characterization, we would also support the proposal made by Ireland to replace lethal force by force, and we would have a preference towards this term rather than the term lethal force. In paragraph 25a, we support the UK proposed language circumstances of engagement.
I thank the distinguished representative of Denmark for stating the position of her delegation. So far, I have no further requests for the floor. Does that at least I don't identify them. Does that reflect reality? I think so. I see no plaques going up. If that's the case, then I would-- yeah. I see one more request from the ROK. You have the floor, sir. And China? Yes. Okay, so we have ROK to be followed by China. You have the floor, sir.
Thank you, Chair, for giving me the floor. We just want to support the Singaporean suggestion at the end of 25A, taking into account the anticipated circumstances of the tech, which I think is balanced and also practical way to go forward. Thanks, Chair.
I thank the distinguished representative of the R.K. for reflecting on proposals made. Next on my list is the distinguished representative of China to be followed by Greece. But let me first pass the floor to the distinguished representative of China. You have the floor.
Thank you, Mr. Chairman. The Chinese delegation would like to take the floor again with regard to the issues apart from the lethality. First of all, with regard to the moment of engagement, we are all aware that ever since the First World War, there are weapons that can be activated passively. In this case, the operator cannot determine before engagement the exact moment of engagement. Therefore, in 25, if we add the moment of engagement, this will dilute considerably the definition of laws and will include a large number of non-autonomous weapons. China has noted that during the statement of India, the India delegation, they also expressed the same idea. We can accept not to delete the expression of the moment of engagement. However, This should not be considered as a prerequisite to exclude from the definition. According to different battlefield circumstances, it can be an element to be taken into consideration. China has noted also that the Singaporean delegation has expressed similar ideas. And given this situation, China wishes to suggest that and the moment of engagement be amended or replaced to the moment of engagement may also be taken into account depending on operational context. If a consensus cannot emerge on this, China would prefer delete the moment of engagement. Furthermore, regarding the proposal by the US with regard to the replacement of human by human being and determined by and determined, China has no objections.
I thank the distinguished representative of China for commenting on the text proposals made by other delegations. Next on my list is the distinguished representative of Greece. You have the floor.
Thank you, Mr. President. At the outset, our delegation would like to convey its deep appreciation for all your tireless efforts in paving the way towards a successful and substantial outcome of our deliberations. Mr President, our delegation fully supports the incorporation of the term lethal force into paragraph 25, as lethality is a crucial element from our perspective, as we have already stated in the past. Moreover, we prefer to retain the wording to identify, select and engage targets rather than using the word can. However, we have concerns regarding the use of the term moment of engagement, as we are not certain that this is technically feasible. We therefore support to use the wording, the circumstances of engagement instead. Thank you.
I have one more request from the floor, for the floor, from the distinguished representative of Brazil. Sir, you have the floor.
Thank you, thank you, Chair. I did not wish to prolong this debate, but I would like to comment on some of the proposals that were made after we discussed this paragraph yesterday. In particular, I would like to thank both, again, the UK delegation for making a proposal to refine the language on 25a. which we had originally suggested. As we said yesterday, we are in agreement with the idea of replacing moment with circumstances. Also, the US proposal to replace engagement with attack is acceptable to us. We believe that this might even make the text clearer. However, I would like to address the proposal made by the distinguished delegation of Singapore I'm afraid that the changes proposed would compromise the original idea of this proposal, which is to prevent this exclusion clause to be used to exclude the range of autonomous weapons based on the fact, the single fact that a specific target has been determined, even if the circumstances of this attack or engagement are completely unknown, so unpredictable to the operators. So I'm afraid to say that that formulation does not address those concerns. And these concerns were magnified by interventions that we heard yesterday that confirm that in some quarters this is exactly the understanding that they have. So this reinforces the need for this clarification to be inserted. I thank you.
I thank the distinguished representative of Brazil for reacting to proposals made and comments made by other delegations. I am just looking around to see whether there is any further request for the floor. I do not see that. I think we can now not conclude, but let's say let's pause the discussion on para 25. I already indicated that if you could put the text on again, sorry, that we will look again at the, I think operate together was able to gather a lot of support, especially if taken into in combination with CAN, although there were a few concerns about that, so we have to see how to deal with that. We have, I think, now some versions of the UK and US proposals with regard to changing at the moment of engagement, so we will look into that. And then with lethal force, we have had also suggestion to just have force or described, describe the potential effect that will give some rise to thoughts on my side. And we will, I think somewhere in the afternoon, we will be able to share with you a new version. And that will be then on the on the menu of this night's discussion in an informal format, because I think today we have to clear up the text with regard to the characterisation. So with that, I think we now have to move forward. with regard to the paragraphs that deal with the human element. Then there are a number of paragraphs from 35 to, I think, 38, with at the heart of it paragraph 36. As was pointed out by a number of delegations in informal consultations, they read this in conjunction with what comes before and what follows after. With regard to the suggestions I sent to you perhaps you can put the text on the screen in paragraph 35 the addition of in of quote International law, in particular, prior to IHL, seeks to align the language with paragraph 36, which already refers to both and thereby to set out the general requirement. Paragraph 36 introduces a number of suggested changes based on proposals made during informal consultations and throughout our last formal session, namely in sentence one, I have deleted the reference to the quote principles unquote as these are already addressed in power of 35. I have also inverted the formulation on quote human judgment and control unquote to quote control and human judgment unquote. since a number of delegations have indicated that mentioning human directly before control leads to unclarity and ambiguity. Then sentence two introduces extra clarity on control, which in my opinion is a necessity for reaching consensus across the room. This sentence is formulated as a positive without focusing exclusively on the operational phase and purposefully avoids moral verbs such as must and should. The condition provided that is added to make clear that control, especially indirect control, is only permissible if human beings have determined that compliance with applicable IHL is ensured. In other words, to clarify that human judgment must be effectuated. This, in my view, helps to balance the para by using agreed language while making a clear link between human involvement and direct or indirect control, all the while maintaining an explicit connection with IHL compliance. Let me underline that my suggestions for para 36 constitute a package deal. whereby sentence two is intended to help balance out the suggested amendment to sentence one. Then paras 37 and 38 simply replicate the formulation of the previous para while streamlining the text and again avoiding modal verbs. I am well aware of the still existing disagreement over whether the framing should focus on what is necessary to ensure control and human judgment or what is needed to comply with IHL. In my view, the suggested changes work together to solve that issue. Measures are not linked to either control and human judgment or IHL compliance, but to both, since control and human judgment are necessary to comply with IHL. In sub-paragraph, I deleted the explicit reference to the number of engagements to streamline the text, also as there is no consensus on whether this is needed when we already have a reference to limiting the scale of the operation. The formulation on the autonomous undertaking has consequently also been deleted, as it was added specifically to accommodate the reference to the number of engagements. In subpara Charlie, the reference to, quote, while active in an operational context, unquote, has been removed. This is not intended to imply that parameter modification is problematic, but rather to acknowledge that such self-initiated and approved changes will be problematic and undesirable regardless of whether the loss is active in an operational context. That's it from my side. As again, some background to what you have received earlier as suggestions by the Chair to facilitate our discussion on the basis of the informal exchanges we have had regarding the draft report. I would now like to open the floor for delegations to comment on the proposals as presented by me earlier. Thank you. I have two requests for the floor. Let me first pass the floor to the distinguished representative of Turkey. You have the floor.
Thank you, Chair. Good morning, colleagues. I'm happy to break the ice. I just have a question for clarification, Chair. Should I go ahead and read all our suggestions from 35 to 38 or should we go para by para because I have a rather long intervention? I would appreciate clarification.
I think this is considered to be, I won't use again package, but we have to see the paras in relation to each other. So I think it's useful to have the full blast of your statement now. Please go ahead.
Well, you asked for it, you have it. Paragraph 35. On paragraph 35, we would ask that the paragraph revert to its earlier formulation limiting the applicable framework to international humanitarian law, specifically applicable IHL. The addition of international law, in particular, broadens the scope of this provision in a way that could bring in other bodies of law not designed to regulate the conduct of hostilities. The addition proposed on 11 August would take us beyond IHL into areas of international law whose application to the subject matter has not been the focus of this group's mandate. On paragraph 36, we can work with the recognition that control may be exercised directly or indirectly, and that compliance is assessed on the basis of the anticipated effects of the attack. This provides useful flexibility for systems relying on decision support and navigation algorithms. We would, however, ask that the sentence, this does not imply constant direct human control present in the earlier text to be restored. This assurance matters to a number of delegations, including mine, and its absence changes the balance of the paragraph. We would also recall, for the record, that human control over the use of force is not in itself an established rule of customary international humanitarian law. A customary rule requires both consistent state practice and the corresponding sense of legal obligation, opinio juris, and in a field where practice is still forming, we do not believe that threshold has yet been reached for a freestanding human control requirement. We raise this not to diminish the practical value of human control as a matter of policy on which we are in full agreement, but to explain our preference that this paragraph continue to be framed in terms that capture common understanding rather than as a codified legal standard. On paragraph 37, we're concerned by the shift away from recommendatory language. The exercise of human judgment and control should continue to be framed with should rather than as an unconditional description of what this exercise entails. We ask for the restoration of should entail in place of the current formulation. As noted under paragraph 36, we do not consider human control over the use of force to be at this stage an established requirement of customary IHL. The same consideration applies here and counsels against treating this exercise as a fixed legal standard in a text that is not intended to create new obligations. We would also suggest supplementing the list of factors with scope of use and circumstances alongside operational context and the characteristics and capabilities of the weapon system to make the list more sensitive to the wide range of context in which these systems operate. Paragraph 38. We welcome the shift toward illustrative language in the chapeau of paragraph 38, namely include inter alia, which we understand to mean that the measures listed are examples rather than exhaustive or mandatory checklist. We would caution against suggestions to make this formulation mandatory, which would not help us reach consensus. Turning to the specific measures on sub-paragraph b, We support, Chair, your proposal to delete and the number of engagements that laws can undertake autonomously. This criterion sits uneasily with certain operational concepts, including coordinated multi-platform operations, where the number of engagements is not a meaningful or practicable metric to fix in advance. Last but not least on sub-paragraph c. We ask for the restoration of the qualifier while active in an operational context. Its removal would extend the requirement for human authorization of parameter modification across the full life cycle of the system, rather than limiting it to the operational phase. This has real consequences for systems that must retain a measure of autonomy to complete their mission or defend themselves when communications are degraded or under electronic warfare or cyber conditions. Alternatively, we could support language providing that parameters are not substantially modified unless such modification or the subsequent use of the modified system is executed or authorized by a human. allowing authorization to attach to the subsequent use of a modified system rather than requiring prior authorization in every instance. This would better accommodate periods of degraded communications, electronic warfare interference or cyber attack in which systems must retain self-defense, safe navigation and mission completion autonomy. Thank you.
I thank the distinguished representative of Turkey for sharing the views of her delegation with regard to the suggestions put forward by the Chair. Next on my list is the European Union. You have the floor, sir.
Thank you, Chair. Since it is the first time we are taking the floor, we would like to thank you for your efforts and excellent leadership in chairing this group. The EU would like to reiterate its support to the following elements contained in the version of the rolling text. Compliance with international humanitarian law. The development and use of emerging technologies in the area of laws must be in compliance with international law, in particular IHL, and we support this reference. Scope for prohibitions in the set of elements should include laws which are of nature to cause superfluous injury or unnecessary suffering, which are inherently indiscriminate. those whose effects in attack cannot be anticipated and limited as required by IHL in the circumstances of their use, and those which are incapable of being used in accordance with IHL, such as the principle of distinctions and proportionality, as well as the obligation to take feasible precautions in attack. Human judgment and control with regard to laws are needed to uphold compliance with international law, in particular IHL. Human judgement and control must be retained during the whole life cycle of the weapon system. We support the inclusion of the set of measures to ensure that such judgement and control are retained. Thank you, Mr. Chair.
I thank the distinguished representative of the European Union for sharing the views of the European Union. Next on my list is the distinguished representative of Stop Killer Robots. You have the floor.
Thank you, Chair. Regarding paragraphs 35 through 38, whatever formulation states settle on to unpack the overall concept that we would refer to as meaningful human control, they must ensure that control and judgment are recognized as activities undertaken by people and that control is exercised by humans. As the current text and the Chair's proposal clarify, this does not mean constant direct control over systems. The role of people and the relationship and interaction with systems has been the central aspect of discussions and must not be eroded or lost. We welcome the Chair's proposal. The reference to appropriate level of human judgment and control has been removed. The level of judgment and control should always be the same, which is to say sufficient, but the measures taken to do this may vary. The subparagraphs a through d remain crucial to the value of the text overall and must be retained. They elaborate aspects of ensuring responsibility and accountability, ensuring that the operation of systems are sufficiently limited in practice and that systems are sufficiently understood. These are all key aspects of meaningful human control. In this regard, we regret that the subparagraph on limiting the operation of autonomous weapon systems to a defined perimeter and restricted use to military objectives by nature that was included in the previous versions of the text has been removed. We are also disappointed to see the suggestion to remove the reference to the number of engagements in subparagraph.
B in the chair's proposal.
We struggle to see why limiting the number of engagements a system can undertake autonomously would not be a consideration for states wishing to uphold the law. A system able to undertake an unlimited number of engagements seem neither lawful or acceptable. Thank you, Chair.
Hi, thank the distinguished representative of Stop Killer Robots for sharing the views of a delegation. Next on my list is the United States to be followed by Portugal, but now first to the distinguished representative of the United States.
Thank you, Chair. Good morning, colleagues. Chair, we're going to follow your guidance and dive right into the text, but we will limit this intervention to paragraphs 35 through the chapeau of paragraph 38 and would plan to address the specific measures in 38 later. I think from our perspective, how these four paragraphs work together will have an impact on how we look at the measures, and so we would like to focus on on the paragraphs themselves at this point. Before doing this, I do want to appreciate the constructive spirit of all of the delegations during our discussions in March on this issue, as well as in the engagements we've had in the intercessional period. These are difficult issues on which delegations have principled and longstanding positions, but I do recognize how hard we and other delegations are working to find the landing zone and we just want to appreciate those efforts. We also appreciate the point just made by the delegate from Turkey who observed that these are concepts, concepts of control and judgment that do not have status as a matter of customary international law. I think for us that's always been an important starting point in this conversation. And it's precisely because they don't have any sort of legal status that it's very important for us to have a shared understanding of how these terms are being used. So, Chair, moving on to the text, for paragraph 35, the previous version of paragraph 35 opened with a reference to applicable IHL, but the new text broadens the framing of the paragraph to international law as a whole. That previous framing, the more focused framing on IHL was acceptable to our delegation because for us it appropriately related the ideas of control and human judgments to the rules governing attacks, distinction, proportionality, and precautions in attack. But the new framing suggests that concepts of control and human judgment are derived from other bodies of international law. And our concern there is that that is perhaps an effort to indicate that they have some legal status from this broader body of law. So in articulating measures to address laws and concepts like control and human judgment, the United States believes it is critical for the group to frame such concepts and measures under the rules of international law applicable to the use of laws and not to the entire body of international law, most of which, or even international humanitarian law, most of which is not applicable to the use of laws. Any use of laws will necessarily be an attack under IHL and subject to all rules and principles that are applicable to attacks, including distinction, proportionality, and precautions in attack. So to have a more focused understanding of how these rules work, we would recommend revising this text to read laws must only be used in accordance with the rules of international law applicable in armed attack. Sorry, applicable in armed conflict. The rules of international law applicable in armed conflict, in particular, the principles and requirements of distinction, proportionality and precautions in attack. The phrase rules of international law applicable in armed conflict is used, for example, in the CCW preamble. And it's also a defined term in Article 2B of the Additional Protocol I to the 1949 Geneva Conventions. And so we hope that because this phrase has a pedigree that is traceable both to the CCW and to the Geneva Conventions that this might be a path forward on this point. Regarding paragraphs 36, 37, and 38, for the United States, control and human judgment are components of IHL compliance. They are distinct but related concepts, but they are not and should not be legal rules or standards. This is why we cannot support articulating a control and judgment standard emanating from IHL that requires a set of measures, but we can support explaining how control and human judgment are part of the implementation of IHL rules governing attacks and elaborating measures related to control and judgment that help implement these IHL rules. So in that respect, Chair, we do welcome the suggestion to change human judgment and control to control and human judgment. As we've pointed out in the past, the group has never agreed on the terminology of human control. And using simply the term control better reflects the past consensus conclusions of the group. In particular, it is essential to reflect the ordinary meaning of the word control. and the fact that existing IHL rules have permitted weapons systems with autonomous features or functions. Automated measures of control have been incorporated into weapons systems for decades and have given effect to human judgments about what is required to comply with applicable IHL. So regarding the second sentence of paragraph 36, we continue to have legal concerns with how the second-half of the sentence indicates that the exercise of indirect control is conditioned on a specific determination. We agree that IHL rules on attack require human beings to make certain judgments in good faith, but we do not think a specific determination of legal compliance is a requirement of existing law nor always practically possible. So After suggesting where we have some concerns with paragraph 36, I would like to try to be constructive in identifying what we think could be reflected here. We can elaborate how control and human judgments are needed under the IHL rules applicable to the use of laws, the rules on attacks, including distinction, proportionality, and precautions in attack. As the GGE has already agreed, these rules require that human beings make certain judgments in good faith-based on the information available at the time. In particular, human beings need to exercise judgment to determine what the rules require in the use of laws. Human beings also need to give effect to those judgments and give effect through those judgments through control directly or indirectly. we see such controls involving measures including measures that can be taken before the use of law so these are all elements chair that we think could could provide us a way forward on paragraph 36 for paragraph 37 as we've said we we don't accept control and human judgment as a standard but we do agree that the factors that are subject of this sentence can inform good faith decision making under the IHL rules governing attacks. Considering these factors is an exercise of human judgment with regard to laws, and these factors should be considered when human beings consider what measures of control need to be implemented under IHL rules. The factors are particularly relevant to whether a measure is feasible and whether it is anticipated to limit effects of attack as necessary under these rules. So These are all elements that we think can be reflected in paragraph 37. Our specific suggestion would be to phrase paragraph 37 as follows. In exercising human judgment to determine what measures of control with regard to laws are feasible and anticipated to limit the effects in attack as necessary under these rules, a range of factors should be considered in good faith-based on the information available at the time. including the operational context human capabilities and limitations and the characteristics and capabilities of the weapon system as a whole um the phrase considered in good faith-based on the information available at the time draws from paragraph 17f of the group's 2019 report and helps provide context for the way ways the the factors should be considered and let me just say chair that here our delegation has raised concern in the past with the the sort of phrase context appropriate. We have said throughout the discussions that if we want to elaborate what we mean by this, it would be something that we would be open to and would be useful doing. And we actually see paragraph 37 as an opportunity to do just that. And so we think it's going in the right direction. And we would like to use this paragraph as an opportunity to elaborate what many delegations have been trying to capture with that phrase. So, I won't address the specific measures in paragraph 38. I would observe that the group needs a common understanding of whether the measures listed in 38 are options to draw from as the state deems appropriate. in light of its domestic context for IHL implementation, or whether the measures listed in paragraph 38 are measures that are always necessary for every state to comply with IHL obligations. And we've heard of a range of views on that question, and it will be really important for us to try to find a common framing for the subparagraphs. To be clear, the United States has been operating under the assumption that it would be clearer and simpler for the group to articulate a longer list of measures that support IHL implementation, recognizing that what is needed for IHL implementation will often be quite different from state to state because states have different IHL obligations and oftentimes very different domestic context for implementation. But we are hearing in the room that many delegations do want to articulate measures that are always necessary and if we're going to go down that road which is not our preference but if if we do end up going down that road I think for us it would be critical for that chapeau on the measures to be clearly connected to the IHO rules on attacks distinction proportionality and precautions and attack the measures listed in 38 would be those measures that are always necessary for the implementation of these rules It's difficult for our delegation to see how what's currently reflected in some paragraph C or D of paragraph 38 would always be necessary to implement these IHL rules on attack. So, Chair, just to conclude, I know that our remarks this morning have been lengthy, but I did want to end again with the thought of how important it is that these paragraphs work together. paragraphs 35 through 38 are connected and the language we use in one paragraph impacts how we read and understand the rest of the paragraphs as other delegations consider the proposals we've set forward set forth we hope that you will keep the overall picture of these paragraphs you and other delegations will keep the overall picture of these paragraphs in mind thank you chair
Let me thank the distinguished representative of the United States for sharing with us the views of his delegation and making proposals. And yes, indeed, we have for good reason considered these paras together because indeed they interact with each other and if we will break changes to certain paras that would also affect then the other paras. So we have to look at this text regarding human intervention, regarding the issue of controlling human judgment as a whole, I would say. Next on my list is Portugal to be followed by Cuba and Ireland. Let me first pass the floor to the distinguished representative of Portugal. You have the floor.
Thank you, Chair. First of all, and since this is the first time I'm intervening, I would like to thank you and your team for all the work carried out so far. We would like also to highlight the constructive discussions, particularly on this set of paragraphs that indeed, as you say, form a package. We are very flexible and could go along with the text as is, without any further change. However, listening to previous speakers, we would also like to contribute to this discussion with the following remarks. First and foremost, we welcome the use of the expression applicable international law, in particular IHL. Because IHL, of course, cannot be applied in isolation. There are several sub-domains of international law that also apply, from international criminal law to the law of treaties, from the law on state responsibility to human rights law that also applies in armed conflict. And we understand the proposal just made by the United States, which is reasonable in our view. However, we believe that the word applicable international law already addresses the issue that was raised by the United States without preempting which norms will apply in a specific situation. So we really would prefer and support the use of the expression applicable international law, in particular IHL, as proposed. On the control and human judgment, we would actually prefer to return to the expression human judgment and control. because control needs to be human as well. And I think line four of paragraph 36, that's, that's, that, that, and I quote, says, human beings exercise such control directly and directly, end of quote, it already captures this idea, the idea that it's not just the judgment that needs to be human control also, is, needs to be human, and this is in our view, important. However, using this expression, control and human judgment could be interpreted as control could be exercised without a human in the loop. Finally, on paragraph 38, we would also welcome the retaining of the expression, and I quote, and the number of engagements that laws can undertake autonomously." In our view, this is an element that is relevant for determining the compliance or actually the gravity of the non-compliance with IHL. I thank you.
I would like to thank the distinguished representative of Portugal for commenting on the text and making proposals. Next on my list is Cuba, to be followed by Ireland, Switzerland and Sweden. Let me now first pass the floor to the distinguished representative of Cuba.
Thank you, Chairperson. We welcome, as on previous occasions, this proposal by yourself. in order to seek the necessary convergences. We have a number of proposals. Some of them have been broadly discussed already within the group. The main focus is that we're in a paragraph that is perhaps critical in order, or a group of paragraphs which is critical in order to find consensus on this document and thus delegation like certain others is concerned. There's a conceptual concern to what degree can control be exercised if there is not first of all human judgment on actions and decisions to be taken because if there isn't human judgment then control is not effective, it's not efficient, and thus we need to look at the wording, human judgment and control. We need to come back to that. Likewise, in paragraph 36, we would propose a wording which would be context appropriate human judgment and control with regard to laws are needed.
English times the effective capacity for human intervention, correction or deactivation of the system.
Additionally, in paragraph 38C, we would have a minor modification to make,
namely the parameters English.
And as additional proposals, we find it a bit repetitive and perhaps unnecessary in this very specific context. In paragraph 38 we would propose perhaps deleting the sentence sorry I'm just looking for it because my computer is giving me some problems Mr Speaker, coming to the proposal to eliminate the phrase 'while', we believe that the proposal is in line with some of the concerns already expressed by certain delegations. But we're seeking this minor amendment in order to bring this text into line with the discussions that we've had on how we should express graduality or stages to achieve effective human control. I thank you.
I thank the distinguished representative of Cuba for commenting on the text and the suggestions from the Chair. Next on my list is Ireland, to be followed by Switzerland, but let me first pass the floor to the distinguished representative of Ireland. Sir, you have the floor.
Many thanks, Chair, for all the continued work here and the work of your team and the ISU as well. As you are no doubt aware, we are more comfortable with the language of your 50 June rolling text that you had here. And in informal consultations in the lead up to this final session, we noted that the inversion of human judgment and control without sufficient strengthening elsewhere could open the door to systems that may not allow for human judgment and accountability necessary to comply with IHL. As we said during previous consultations, a majority of delegations have moved a very great distance to accommodate the concern of some delegations, And we are willing to keep working to find a landing zone around these paragraphs. But we need to ensure that we avoid a text that becomes permissive of systems that do not allow humans to exercise the judgment that IHL requires. With that in mind, in order to maintain a substantive report that would live up to the group's mandate, we want to propose a change to the suggested language you have on the screen on paragraph 36. We do this with the knowledge that many high contracting parties may not have previously heard our suggestions in recent weeks. I will try to read this out slower than most Irish would speak and hopefully the interpreters will be able to process this as well. So the paragraph for 36 would read as follows: To uphold compliance with international law, in particular IHL, including the principles of and requirements of distinction, proportionality and precautions in attack. Control and human judgment with regard to laws are needed. This does not imply constant direct human control. Human beings may exercise control directly or indirectly. Full stop. Regardless of how control is exercised, human beings must retain the judgment necessary to determine that the effects of the use of laws will comply with IHL in the circumstances of their use and must retain the capacity to give effect to that judgment throughout the operation of the system." End quote. I can read this out again if I see nodding heads. I'll read it out one more time for the interpreters and for the team. Quote, "To uphold compliance with international law, in particular IHL, including the principles and requirements of distinction, proportionality and precautions in attack, control and human judgment with regard to laws are needed. This does not imply constant direct human control, semicolon. Human beings may exercise control directly or indirectly, full stop. Regardless of how control is exercised, human beings must retain the judgment necessary to determine that the effects of the use of laws will comply with IHL in the circumstances of their use and must retain the capacity to give effect to that judgment throughout the operation of the system." Chair, I'd just like to explain our proposal and how we believe it addresses three issues. The first is retained capacity. Humans must retain not only the judgment necessary to determine the effects will comply with IHL, but the capacity to give effect to that judgment throughout the operation of the system. To be clear, this is not through constant direct human control. It is retained capacity and it is a standard already met in existing state practice. What it prevents is a single pre-activation determination standing in for human judgment indefinitely. Second is temporal scope. Our concern is that as autonomous operation extends in time, the gap widens between the moment of human judgment and the application of force. A determination made once before activation becomes progressively less meaningful. We would want judgment tied to the circumstances of use, language this group has already used in paragraph 32. Subparagraph 38 should also retain the ability to limit the number of autonomous engagements. That is one of the few concrete constraints on temporal scope anywhere in this text. Third, the chapeau of paragraph 38 must remain mandatory. For us, these measures are not optional. They are what makes compliance possible for systems of this kind. If the inversion is accepted, as we are working with now, and the chapeau is softened, and the measures become illustrative, then the text loses its core added value and we are left with restating IHL. We also support suggestion made in previous informal consultations by Austria on paragraph 37, namely changing entails to requires and introducing language on the protection of civilians. both of which would help strengthen the overall package in our view. We also, like other delegations and an eloquent put forward by Portugal and others, want to see the retention of the reference to international law in paragraph 35. So in sum, our proposal for these paragraphs is to retain international law in paragraph 35, a modification of paragraph 36, the language we have just outlined and we will also send through to you. paragraph 37 would be maintained along with the Austrian suggestion and finally the chapeau of paragraph 38 would remain mandatory. Chair, in line with your previous utterances, we echo the importance of the work we are doing this week in fulfilling the mandate of the GGE. I think for my delegation it is also important to stress that the purpose of the CCW and our deliberations here is not to merely repeat previous GGE report language nor merely to restate existing rules. It is to explain how these rules operate when the technology changes, The convention exists precisely because general IHL principles require elaboration as new means and methods emerge. We will share all the following in writing and we are happy to engage on the proposals as they develop. Thank you.
I thank the distinguished representative of Ireland for commenting on the chair's suggestions and for making proposals in relation to those proposals and reacting to the remarks of other delegations. Next on my list is Switzerland, to be followed by Sweden and Italy, but let me first pass the floor to the distinguished representative of Switzerland. You have the floor.
Mr. Chair, we can broadly support your text as it was circulated in the rolling text in paragraphs 35 and following, but see that some changes are still needed to bring us over the finish line. Having said that, while we are willing to show flexibility, we are also mindful of the cumulative effects of such changes taken together, small adjustments could upset the fine balance achieved over the past months, if not years. The following general points are important for us. On judgment and control, it is important that neither judgment control nor the taking of concrete measures appears optional. The text should be clear that human judgment is always required to assess whether the use of laws, including their anticipated effects, would comply with IHL in the circumstances of their use. On control. We can accept language saying that control may be direct or indirect and that it may be exercised through measures taken before or during use. We understand that a text agreeable in this room would not imply constant direct human supervision. Regardless of the form it takes, control must ensure that relevant human judgment has a practical effect on the system's operation, including by limiting the system's anticipated effects. On measures, the text must be clear that a prior assessment by human beings alone is not sufficient. The necessary technical and operational measures must actually be implemented. paragraph 36 to 38 must therefore remain clearly linked. We have also heard with great interest a proposal just made by Ireland and we are happy to study it in more detail as a basis. Mr. Chair, let me also conclude with some remarks on paragraph 35. We fully support your suggestion to keep a broad formulation. referring to applicable international law. However, we also heard the US proposal on international law applicable to armed conflict, a formulation indeed contained in the preamble of the CCW, which could be a way forward, perhaps without rules as beyond the rules and the international agreement, it encompasses, in our view, also the generally recognised principles and rules of international law which are applicable to armed conflict. Thank you, Mr. Chair.
I thank the distinguished representative of Switzerland for commenting on the text and proposals made by other delegations. Next on my list is Sweden, to be followed by Italy. Then I have Human Rights Watch and the Republic of Korea. But let me first pass the floor to the distinguished representative of Sweden. You have the floor.
First, we would want to emphasize that for many years our discussions demonstrated broad support for meaningful human control, and for the last couple of years, a vast support for context-appropriate human judgment and control. We would like to recall that for us and for many other states, the latter notion was already a compromise. The vast support signifies that it is essential to keep humans at the center of the development and use of laws, and that this needs to be clearly set out in the set of elements. Sweden support human judgment and control. The necessary control will depend on what is required to comply with IHL in the operational context, And we consider that both the draft report and the additional suggestions expresses adequately this need for adaptability. We welcome the Chair's suggestions and other proposals that explains that humans can exercise control directly or indirectly, including through technical systems. We appreciate the constructive discussions here today and we remain open to suggestions that captures the core of this notion and what it entails. We believe that we can find consensus. I thank you.
I thank the distinguished representative of Sweden for commenting on the text and on the development of the text. Next on my list is Italy, to be followed by Human Rights Watch and the Republic of Korea. But let me first pass the floor to the distinguished representative of Italy. Madam, you have the floor.
Thank you, Chair, for giving me the floor and allowing me to comment on these paragraphs. With regard to paragraph 36, we note the suggested inversion from human judgment and control to control and human judgment to allow for indirect forms of control while still preserving opportunities for human judgment and guaranteeing accountability. Although we would have preferred the simpler, concise wording you used in the earlier draft of the rolling text, we are prepared to show flexibility and accept this change as part of a package together with the second sentence of this paragraph. Regarding paragraph 38A, we support the current wording focused on the responsible chain of command and control, which determines and evaluates on the basis of a legal assessment, the effects of a target identification, selection, and engagement, taking into account the operational context, as well as the reference to ethical considerations, drawing inspiration from the guiding principles. We also support the sub paragraph B, which establishes the responsibility of the relevant entity to limit those effects through appropriate operational constraint, including inter alia restriction to the scale of operation, types of targets, duration and geographical scope. We favored the removal of the reference to the number of engagements, as in our assessment, this criterion would not constitute an effective form of control and could, in practice, result in the proliferation of systems. Regarding subparagraph c, we deem it relevant as it relates to human intervention in the event of significant changes to mission parameters, limiting the use of AI systems which self-learn after activation and could independently modify human-entered mission parameters, raising the risk of unintended engagements. Finally, we encourage retention of subparagraph, which provides for the inclusion of automatic manual deactivation systems for additional protection. Allow me to end this intervention, Chair, by underlining that we regard these paragraphs as of paramount importance, and we hope that we can narrow the remaining differences in the conference room by the end of this week. Thank you, Chair.
I thank the distinguished representative of Italy for sharing the views of her delegation and commenting on proposals made. Next on my list is Human Rights Watch to be followed by the Republic of Korea. But let me first pass the floor to the distinguished representative of Human Rights Watch. You have the floor.
Human Rights Watch welcomes the years of work that you and the Group of Governmental Experts have put into discussing elements of a new instrument on lethal autonomous weapons systems. The draft text structure of prohibitions and regulations with an emphasis on the role of human control, the role of humans in the use of force provides a valuable framework for addressing many of the concerns raised by autonomous weapons systems. We urge states this week to adopt the draft text in either its current form or stronger one. Anything less would undo the progress that they have made over 12 years of discussions under the CCW's auspices and the three years of drafting under this mandate. Among the key elements to maintain is paragraph 36, which recognizes the need for human control and judgment. The paragraph enumerates several components of that concept, including oversight by a responsible chain of command to ensure that a human can assess compliance with legal obligations and ethical considerations. Operational restrictions, such as those on scale, type of target, time, space, and we believe should be included number of engagements. assurances that only a human can change this system's target admission parameters and the ability to deactivate or neutralize the autonomous weapon system. We recommend that these points be retained and while many of the revisions proposed in the chair's paper effectively fine tune the language, we also strongly recommend that the word human is linked to both judgment and control. We also welcome the inclusion in paragraph 35 of the language with applicable international law, in particular IHL, because it recognizes there are multiple areas of relevant international law, including international human rights law, and is consistent with earlier CCW documents on laws, such as the 2024 Guiding Principles. There are additional elements of the text that are important to preserve, such as references to preventing and mitigating bias, provisions to ensure that autonomous weapon systems are predictable, reliable, traceable, and explainable, and measures to promote accountability. There are other details that could be refined and some areas that should be strengthened, notably the prohibitions should extend to autonomous weapon systems that target people, and there should be other references to international human rights law, but these changes can wait for negotiations. In the meantime, high contracting parties should hold their ground. embrace their accomplishments by adopting the draft text and take a concrete step forward at the review conference by adopting a mandate to start negotiations. Thank you.
Let me thank the distinguished representative of Human Rights Watch for stating the positions of her delegation. and the further development of the text. Next on my list is the distinguished representative of the Republic of Korea to be followed by New Zealand. But let me first pass the floor to distinguished representative of the Republic of Korea. You have the floor.
Thank you, Chair, for giving me the floor. I'll keep my intervention short. First, my delegation welcomes the inversion of human judgment and control to control and human judgment, which makes us closer to consensus. Second, we share the similar sentiments with the US and other delegations regarding the nature of measures in paragraph 38. We view them as illustrative example of measures needed to implement what is required by applicable IHL. They are important and practical measures, but we do not see them-- all measures should be applied in all circumstances. In other words, what measures are feasible depend on the range of factors enumerated in paragraph 37. The current Chair's text which says include inter alia in paragraph 38 supports our understanding. Third, we welcome the deletion of number of engagement in sub-paragraph b of paragraph 38 and understand that such measures can be included in a more broad term, the scale of operation in the same paragraph. Also, we concur with our colleague from Switzerland that the language from CCW preamble, namely rules of international law applicable in armed conflict could be a good way forward. Lastly, regarding the Irish proposal, although we appreciate Ireland's constructive efforts to introduce a compromise text, at this stage we would like to focus on the Chair's additional suggestions as the new proposal introduced concepts such as retention that we need more time to understand. Thank you, Chair.
Let me thank the distinguished representative of the Republic of Korea for stating the positions of his delegation and commenting on the text proposals of other delegations. Next on my list is New Zealand, to be followed by Egypt and Germany, but first the distinguished representative of New Zealand. You have the floor, sir.
Thank you, Chair. proposition for New Zealand across paragraphs 35 to 38 is that human judgment and control with regards to laws are needed to uphold compliance with IHL. For us, it's essential that the text retain both the connection to applicable IHL and the practical measures necessary to make human judgment and control effective. Provided that core relationship is preserved across paragraphs 35 to 38, we are open to considering different proposals that will help us achieve consensus. We note the concrete proposal from Ireland made this morning, which we can see as a carefully crafted attempt to find a balance between various views that have been expressed during previous consultations, and we are studying it carefully. On the inversion of the language in paragraph 36 to control and human judgment, we think there is potential that reversing the order risks creating ambiguity about the nature of the control required. and placing greater emphasis on supervisory control when human judgment is also needed in making substantive legal and operational decisions. So we continue to prefer the original formulation. Nevertheless, we can consider the inversion if it is part of a broader package that gets us to consensus on this text, and we retain those essential elements I mentioned in opening. We do have a concern with the formulation of the second sentence in paragraph 36, in particular the phrase, "provided they determine on the basis of the anticipated effects of the attack that such use will comply with applicable IHL." To us, this could be read as suggesting that human accountability, judgment, and control are satisfied simply by making a finding of legal compliance. In our view, an assessment of IHL compliance is necessary, but it does not by itself demonstrate that the system is subject to the necessary human judgment and control. We would propose a revision that would read, provided they determine on the basis of the anticipated effects of the attack that such use will be carried out in compliance with applicable IHL. The words carried out in compliance we think would provide a clearer connection between the human determination and the manner in which the use is to be conducted. This is a limited clarification and of course the practical measures identified in paragraph 38 remain essential. On paragraph 37, we would also support changing entails to requires, and we would support including a reference to presence of civilians as a factor for consideration. On paragraph 38, we have one suggestion related to subparagraph c, which we also made in the informals, and that is we would suggest that substantially modified could be replaced with materially modified. to ensure we capture smaller changes that may not by the ordinary definition of substantial be substantial, but nevertheless have a material effect. And I'll leave my comments there. Thank you, Chair.
I thank the distinguished representative of New Zealand for commenting on text proposals and for offering Next on my list is Egypt, to be followed by Germany, South Africa and Chile. But let me first pass the floor to the distinguished representative of Egypt. You have the floor.
Thank you, Mr. Chair, for giving me the floor. As you are proceeding through this text paragraph by paragraph, we would, for the sake of coherence, deliver our comments together as a single set of remarks, especially that most of these remarks relate to paragraph 38. We hope that that will assist the group in seeing the full picture of our concern at once, and we thank you for your indulgence in that regard. First, on paragraphs 36 to 38, we understand and appreciate that the ordering to control and human judgement is offered in the spirit of compromise, and we recognise the effort behind it. We would nonetheless like to share our concern about the effect of this particular formulation beyond its wording alone. Throughout this group's work, human judgment has served as the conceptual anchor from which control drives its meaning. It is judgment that grounds the determination of IHL compliance and control that gives that determination effect. Placing control first risks into inverting that relationship. It could be read as treating control as the operative starting point, with judgment following as a secondary or confirmatory step, rather than the basis on which control is exercised in the first place. In practice, this creates room for a system to be considered compliant on the basis of technical control mechanisms alone. With the substantive exercise of human judgement, the element that IHL accountability actually depends on is reduced to an afterthought. We would recall that in March, we have already proposed language addressing this precise point. To ensure compliance with IHL, human judgement and control throughout the life cycle of laws is required. That formulation was not put forward as a national preference alone. It was designed as a bridge between perspectives that viewed human judgment and control as the hard-won mandatory core of this text and those who viewed it as a mechanism for achieving IHL compliance rather than a standalone legal norm. By placing the IHL obligation first and framing judgment and control as the instrument through which that obligation is fulfilled, The formulation preserved the concept of mandatory character while directly answering the concern that it is not to be read as an independent rule. It also preserved the correct sequencing of judgment before control and integrated life cycle application directly into the operative requirement, addressing in a single sentence three of the concerns we have just raised. We would therefore ask to consider reverting to that formulation rather than developing new language that reintroduces gaps that it had already closed. Secondly, we note the deletion in paragraph 38 of the limit on the number of engagements that LOS can undertake autonomously. That was one of the few quantifiable and verifiable safeguards in the text. Without it, the measures in paragraph 38 become harder to operationalise and easier to satisfy in form without substance. It would be better to restore this element. Third, paragraph 38 continues to lack an explicit addressee and obligation verb. Without a named subject and an obligation verb, the measures read as a list of options rather than expected practice, leaving States free to treat them as aspirational. It would be better for states to be named as the subject using shall or should ensure. Finally, we note the continued absence of a divergent paragraph notwithstanding the support recorded for it in the chair's own summary for the first 2026 session. Its continued absence leaves a recognized risk area unaddressed despite the group's own record of support for closing that gap. It would be better for that element to be restored. On a positive note, we welcome the restoration of article 36 consistent trigger language at paragraph 38 and the addition of predictable at paragraph 40. I will close with a broader reflection. Human judgment and control are not one element among many in this text. It is the core around which the entire set of elements is organized and the principal assurance that this group can offer that IHL compliance remains meaningful in the face of increasing autonomy. If the language governing this concept is allowed to weaken, the weakening will not remain contained to a handful of paragraphs. it will weaken the credibility and the value of the document as a whole. We say this in a constructive spirit as we wish to see a successful outcome for this group's work. But success in our view cannot be measured only by whether the group reaches consensus. It must also be measured by what the consensus actually protects. A text that commands that commands agreement but no longer safeguards human judgement and control would not, in our assessment, represent progress. In such a case, we would respectfully submit that it is preferable for us to acknowledge that further work is needed, rather than to conclude with language that has been emptied of the substance that justified this exercise in the first place. I thank you, Mr Chair.
I thank the distinguished representative of Egypt for sharing the comments and reflections of his delegation. Next on my list is Germany, to be followed by South Africa and Chile. But let me first pass the floor to the distinguished representative of Germany. You have the floor.
Thank you, Chair. These paragraphs are, of course, at the heart of our discussions. As we've underlined in the past, Germany attaches great importance to ensuring that a framework of human judgment and control is in place during the life cycle of a weapon system. In our view, the current formulation in the draft final report has emerged as a compromise in order to try and bring together delegations' view on this concept. However, during the last formal and informal sessions, it has become more and more clear that this compromise does not work for all delegations. We therefore appreciate your efforts and efforts by delegations to rephrase this para in a way that anchors the human element firmly in the text and addresses delegations' comments during the last sessions. Regarding the inversion of the first sentence of para 36 to control and human judgment, we do prefer the current version of the draft report. However, in our view, the content of the text as a whole, and particularly of these paragraphs on the human element, is more important than the specific word order in question. We are still looking at the second sentence of your suggestions on para 36 in more detail. We would like to thank delegations for coming up with additional text proposals to further work towards consensus on this paragraph. We would like to offer a preliminary comment on the Irish proposal, which in our view addresses some of the questions we had regarding the second sentence of these additional suggestions. We see the wording, regardless of how control is exercised, human beings must retain the judgment necessary to determine that the effects of the use of laws will comply with IHL in the circumstances of their use and must return the capacity to give effect to that judgment as a simpler and therefore preferable way to express how judgment and control are related to one another and to IHL and the circumstances of the use. However, perhaps there could be some fine tuning to the last part of the sentence of the proposal, which currently reads as, and must retain the capacity to give effect to that judgment throughout the operation of the system. In our view, the last bit throughout the operation of the system might be read in a way that does imply constant direct control or constant direct oversight throughout the use of the weapon system and therefore could be read in a contradiction to the previous sentence. Also, as we've underlined previously and previous speakers have underlined today, this is not required by IHL. It might be a possibility to shorten the sentence so it would read, regardless of how control is exercised, human beings must retain the judgment necessary to determine that the effects of the use of laws will comply with IHL in the circumstances of their use and must retain the capacity to give effect to that judgment. Regarding 38, we could support the deletion of the number of engagement as this might not be applicable in all cases, but would in our view be covered by restricting the scale of the operation. And regarding 38, we support the deletion of the phrase while active in operational context. We share the understanding that any substantial modification of a target or mission parameters by a loss should be executed or authorised by a human, irrespective of the point at which that modification occurs. Thank you, Chair.
I would like to thank the distinguished representative of Germany for commenting on the text and proposals made by other delegations. Next on my list is South Africa, to be followed by Chile, the Czech Republic, France, India, Mexico, Brazil, and then we have South Africa, Spain, Peru, Austria and the Russian Federation. I have now twice South Africa, so that's minus one then. Let me now pass the floor to the distinguished representative of South Africa. You have the floor.
Thank you very much, Mr. Chair. My delegation would like to make the following comments under these paragraphs 35 to 38. Particularly on paragraph 36, South Africa wishes to express its preference For the previous formulation prior to June 5, referring to human judgment and control, in our view, both concepts remain important and mutually reinforcing. Human judgment preceding control is critical in ensuring informed and responsible decision making, while human control remains an essential safeguard for maintaining accountability and compliance with international law, in particular IHL. Accordingly, South Africa would favor language that explicitly reflects both human judgment and human control in that sequence. Turning to paragraph 38, South Africa supports the principle that significant modifications to target and mission parameters should remain subject to appropriate human responsibility, authorization, and oversight. At the same time, we believe that the text should maintain a degree of flexibility and adaptability to account for the realities of contemporary operational environments. Modern battlefields are dynamic and often require responses to rapidly evolving circumstances. In this regard, South Africa would support language that preserves meaningful human involvement while avoiding overly rigid formulations that may not adequately reflect operational realities. For South Africa, the key consideration remains that any modification or adaptations must continue to be undertaken in a manner consistent with applicable international law, including international humanitarian law. Maintaining compliance with legal obligations is paramount while ensuring that the framework remains practical and implementable. Thank you, Mr. Chair.
i thank the distinguished representative of south africa for sharing the views of her delegation and commenting on proposals next on my list is chile to be followed by czechia but first the distinguished representative
of chile
thank you president on paragraph thirty five we support it as is now in particular the generic reference to applicable international law should remain. Autonomous lethal autonomous weapons systems, kennels, be part of conflict, then it's international rights law applicable and appropriate international law, including police and international security. So explicit mention of IHL is a particular reference within a broader framework. On 36 to 38, 36, we would prefer to keep the second line to keep the principles of distinction, proportionality and precaution. Having said that, Chile could, as a compromise, have control and human judgment should be to be joint read jointly with 3738 thus keeping the link between human judgment control and fulfillment of IHL. This requires an intermediate this is an intermediate standard not requiring direct constant human control. This is required but it could also be within the national purview but as a minimum that it has to be tangible application within the operational system. We could consider positively the Irish proposal made just now and coming back briefly to 35, the reference to force as a compromise we could live with within force but not lethal force. And on this, the applicable legal framework of 36 to 38, we would not be, it's not aligned, and we should not be broadening 38, but rather within each of those paragraphs, We would seek not to extend this since years have been spent on seeking this formulation. Thank you.
I thank the distinguished representative of Chile for sharing the views of her delegation. Next on my list is the Czech Republic, to be followed by France and India. Let me first pass the floor to the distinguished representative of the Czech Republic.
Thank you, Mr. Chair. We would like to thank you for your continued efforts over the recent months to bridge the different views expressed regarding paragraphs 35 to 38. For us, the central issue remains compliance with applicable international humanitarian law. We therefore welcome the affirmation in paragraph 35 that laws must only be used in accordance with applicable international law, in particular IHL, including the principles and requirements of distinction, proportionality, and precautions in attack. In our view, this text provides the appropriate foundation for the following paragraphs. We view control and human judgment as mechanisms that contribute to ensuring such compliance. They should not be understood as creating a new and independent legal standard existing alongside or going beyond the existing rules and obligations of IHL. Rather, they are a means through which compliance with those rules may be achieved in particular circumstances. We therefore welcome the efforts undertaken in the run-up to this final session to reinforce the link between control and human judgment and compliance with applicable legal obligations. In our view, the decisive question is not whether a weapon system incorporates autonomous functions, but whether that system can be used in compliance with IHL in the circumstances of its anticipated use. We also welcome the clarification that human beings may exercise such control directly or indirectly. We understand this wording as recognizing that relevant human judgment may be exercised at different stages of a system's lifecycle, including during legal review, testing and evaluation, and in preparation for a system's deployment. In particular, such judgment may be reflected through defining the system's intended purpose, permissible target profile, geographical and temporal limits, conditions for activation and termination, as well as other technical and operational parameters established before deployment. Therefore, human judgment need not be exercised only immediately prior to each individual engagement, but also throughout the development, preparation, and deployment process. At the same time, however, the reference to direct or indirect control should not be interpreted as permitting the transfer of legal judgment to another autonomous system. While technical systems may identify, recognize, or classify objects on the basis of predefined parameters and available data, they should not be understood as performing the legal assessment required under IHL. The legal assessment of the conditions under which a category of targets may be engaged and the decision to employ the system in a particular operational context must remain with human beings. Human beings likewise remain responsible for the decision to employ the system under the circumstances concerned. Now, turning to paragraph 38, like some other delegations, we consider the measures listed there to be useful examples of how control and human judgment may be exercised in practice. It should nevertheless remain clear that these measures are not intended to constitute a mandatory and cumulative checklist applicable at all times. The appropriateness of any particular measure will necessarily depend on the characteristics of the system concerned, the nature of the operation, and the operational environment in which it is expected to function. In this regard, we welcome the deletion of the number of engagements as a separate parameter. We do not consider the number of engagements taken in isolation to be a meaningful indicator of whether the use of a system complies with IHL. In some operational scenarios, including these against multiple incoming threats, a system may need to conduct a significantly higher number of engagement within a very short period of time. Compliance with IHL should therefore be assessed in light of the circumstances of use and the anticipated effects of the operation, rather than on the basis of a predetermined numerical threshold. Thank you, Mr. Chair.
I thank the distinguished representative of Czechia for commenting on the text. Next on my list is France, to be followed by India and Mexico. But let me first pass the floor to the distinguished representative of France. You have the floor.
Thank you, Chairperson. The question of the human element is central to our work, since we are committed to explaining how IHL applies to autonomous systems. So I would like to explain how my delegation understands this control and why we can support your proposal. IHL requires the command reasonably to anticipate the effect of a weapon, including that these effects be directed against a specific military objective and that incidental damage to protected persons and property be limited. Humans must ensure that their actions and the use of force comply with the rules and principles governing the conduct of hostilities. They must therefore exercise control over the autonomous systems that they deploy. This control is determined on a case-by-case basis, depending on the characteristics of the system, its mission, its field of use, and on the basis of any element that may be relevant to the military command. This control can take a number of forms, from direct or operational human control to more indirect control. And this control can be exercised throughout the life cycle of the weapon system, in particular during its development or deployment. The restrictive measures that make it possible to carry out that control are subject to the context in which they are to be applied. They're not always all necessary. We find all of these elements in paragraphs 35 and 38 as proposed in your draft, and we can therefore support them as they stand. However, we also heard during the informal consultations other proposals, in particular from the Irish delegation that we heard again this morning, and the UK delegation last March. And they contain certain elements which we think are promising. First, the idea that control is carried out by different measures, which was expressed as measures of control. Then the idea of the concept of human judgment is specified. This is dissociated from the way in which the control is carried out. And we can translate it as regardless of how control is exercised. Third, the need for the command to be able to exercise judgment on the effects of the weapon. This is the judgment necessary term. And finally, the idea that the human command acts on the basis of this judgment, and we call that to give effect to that judgment. Thank you.
I thank the distinguished representative of France for commenting on the suggestions circulated and on proposals made by other delegations. Next on my list is India, to be followed by Mexico and Brazil, but I first pass the floor to the distinguished representative of India, Surya Pathak.
Thank you, Mr. Chairperson, and good afternoon to all the colleagues. Mr. Chair, we want to briefly underscore that we are encouraged by the depth and the nature of debate that we have seen over the last one day and today. And we are also happy to see the standard of articulation that the states have exhibited in trying to articulate their interests. and that we are happy to see the constructive engagement of delegations towards a meaningful and substantive outcome of the group. Mr. Chair, before I share my delegation's views on paragraph 35 to 38, which is on the core measures, I want to just briefly underline that we do have some questions, if not concerns, on the characterization aspect, particularly paragraph 25a. We heard a number of views today by several delegations and we want to just emphasize on one of the edits proposed by the Singaporean delegation on taking into account the anticipated circumstances of engagement, which my delegation considers positively, and we're willing to have more discussions on that. And I also want to reiterate that we still want to seek more clarification in terms of how the three verbs designs, programmes and determines applies to a set of potential targets. We are not opposed to it, but we would like to seek more clarification on that aspect. Now on the core measures, paragraph 35 to 38, we have heard very carefully to delegations this morning. We want to just underline that we believe that you have done a remarkable job in trying to lay out what could be a potential landing zone. We also are fully aware that delegations have deeply held and strong views on this aspect in terms of capturing the human element. And despite those deeply held and strong views, you've tried to propose something that covers everybody's essential points. And I would also iterate what the UK delegation had stated on the first day that We should try to distinguish between what is essential on one side and what may be preferable on the other. So we would echo in the comment that they made. And of course we believe that the text that you propose should be the basis of our negotiations and we believe that it should not be reopened by a number of changes, particularly by new conceptual areas which could produce a setback to our discussions. Now textually on para 35, 36 and 37 first. My delegation can support the structure and I'm noting the word structure of the paragraphs and the logical flow between them. We do agree with some of the comments made by delegations including the US, Republic of Korea on replacing the reference to international law with international law applicable to armed conflict which is used in the preamble of paragraph eight of the CCW framework convention. because if this change is not made then we might risk covering other bodies of international law which may not be appropriate to the conduct of armed conflict. So that would be the recommendation from our end. On para 36 and 37 we are at this point I'm happy to report that we are comfortable with the label of control and human judgment which as you pointed out in para 36 is needed to comply with the IHL. We can also agree with the contents in para 37 which lays emphasis to consideration of operational context as well as the characteristics and capabilities of weapon systems as a whole in deciding control and human judgment. We do however note that some delegations have proposed some edits on para 36 with respect to what has been accomplished understanding of this group with respect to good faith. and I want to refer back to the 2019 report and paragraph 17f, which obviously, as you are aware, states that human beings make certain judgments in good faith based on the assessment of information available to them at that time. So we would be happy to see if this text could be improved by those edits that have been proposed. Mr. Chairperson, on para 38, we do have some concerns. We see the measures outlined in para 38 as measures that could support compliance with IHL. We do not view at this stage these measures as necessary in and of themselves. And we also understand that states might implement one or two of them and not all of them in one go. We also agree with your view that this is an illustrative list of measures and we see that you made reference to inter alia and that this is not an exhaustive list. Secondly, on this paragraph, we want to suggest that we may look at the template of the para 22 of the 2023 report, which is the first time when the group actually listed measures. That paragraph was carefully negotiated among States and that had a carefully built understanding about how these measures are viewed by delegations. That paragraph obviously says that these measures are to be, you know, so the formulation which is used is where necessary states should. So that would be a suggestion from our end which I think we can take into account and we will be happy to discuss this further with delegations in trying to correctly enshrine the relationship between measures and the level of control and human judgment that you proposed. Now on the individual measures that you have proposed, My delegation can support the reference to ethical considerations. We feel that this is an important aspect and we can preserve it in the manner you have shared. I also want to propose an edit in subparagraph which we had also shared during the informal consultations, that we would suggest the use of a more neutral term such as adjusting in the place of the word restricting. And in subparagraph My delegation has some concerns, more in the form of questions with regard to the interpretation of the term substantial modification. Our question is that what is the threshold that the term substantial modification implies and what differentiates substantial from non-substantial modification? This is in form of a question and less in form of a concern, but we are willing to discuss this further. And on subparagraph d, We want to just underline that it may not be possible to deactivate or neutralize laws always in all cases and that too in a timely manner. This is a design feature of the weapon system. Therefore, in our view, this measure can be treated in taking into account the characteristics of laws. We also agree with the view that along with deactivation or neutralization laws can be redirected to another target or for another purpose. So the formulation could be kept more open-ended. That would be the suggestion from my delegation and that there could be a variety of ways in which laws could be aborted. So we would prefer formulations which are less prescriptive. And with this, Mr Chairperson, I conclude, but we look forward to thrashing out the concept and the measures in details and in trying to help you out on this, on these paragraphs. Thank you.
I thank the distinguished representative of India for commenting on the text and the proposals made. Next on my list is Mexico, to be followed by Brazil and Spain, but let me first pass the floor to the distinguished representative of Mexico. Madam, you have the floor.
Thank you, Chairperson, and good morning, colleagues. I will try to be brief and address the different paragraphs for four hours. 5. I reiterate that for us, IHL doesn't apply in this tool. And so we welcome the tool, the text proposed by you on the screen now as it stands. On 36, and particularly the discussion on human judgment and control, We have said on a number of occasions the importance of the concept of judgment and human control, that's reasonably clear. So legal and contextual assessments are needed to determine whether an attack is illicit. That must be done by human beings and there need to be effective measures to guarantee the operation of the effects of the system that it remains within the authorized limits authorized by humans and as mentioned during the informals last week the wording is a little problematic for us however in order to resolve the deep divergences on this topic and find a compromise and in response to your call chairperson we can be flexible and accept controlled human judgement, the new wording, as long as it is within the context of the Irish proposal for this paragraph and the following paragraphs. We have reconsidered the Irish proposal and we consider that it provides clarity regarding fulfilment of IHL and constant human control and recognise that the control must be done in the final instance by humans directly or indirectly. Likewise, it is important here to maintain an explicit mention of the requirements and principles of distinction, proportionality and precautions in attack as reflected in prior versions of the text. On paragraph 37, we would like to reflect the word necessary once again in order to express human control in relation to various factors. On 38, we regret the loss of the reference to a limit of actions or engagement and we would like to explore a way of providing clarity on what can be autonomous and what must be limited. while leaving States with the necessary flexibility to determine the adequate limits for these actions. I thank you.
I thank the distinguished representative of Mexico for commenting on the text and proposals made. Next on my list is Brazil, to be followed by Spain, Peru and Austria. Let me first pass the floor to the distinguished representative of Brazil.
Thank you, Mr Chair. At first, my delegation would like to say that we regard this set of paragraphs as absolutely crucial for the work of the GGE. And it's the strength of this section that will determine whether the draft report will represent a meaningful contribution to the objective of regulating autonomous weapons, in our view. That said, we thank you very much for the preparation of the proposals that are in front of us, many of which represent a marked improvement vis-a-vis the June 5th version of the rolling text, which in our view already represented a sound basis for the adoption of our substantive report. Very few remarks on the points that still merit in our view closer consideration. First on paragraph 35, we strongly support the formulation that you have proposed, Mr. Chair. As we heard from other delegations, this open-ended reference to international law would make this report relevant to elements that to other other discussions on laws that go beyond the strictures of the CCW. And as others have mentioned, IHL is not the only body of law that applies to armed conflict. So we would strongly prefer the formulation that you have proposed. As for paragraph 36, as we indicated during the informal consultations, we are concerned with the reversal of the order of the terms control and human judgment for reasons that have been expounded by many other delegations as well. We believe that this change must be considered in terms of the possible overall impact on the section as a whole. Like Switzerland and New Zealand, we underscored that this section should not lend itself to the interpretation that the prior assessment of expected effects is sufficient to ensure compliance with IHL. Therefore, we welcome the proposal made by Ireland, which in our view would contribute to maintaining the balance in the package of formulations regarding human involvement. We attach particular importance to the expression that is contained in the Irish proposal, I quote, to give effect to that judgment, which in our view corrects any possible misconceptions about what is necessary in this field. Turning to paragraph 37, we support the proposal that was made by Austria during the informal consultations to include the reference to the protection of civilians in that language. And turning to paragraph 38, like many other delegations before us, we underscore the need to retain a clear connection between the references to human judgment and control that are present in paragraph 36 and the measures contained in this paragraph 38. In our view, these measures are what give effect to the formulation human judgment and control and provide objective elements for the implementation. So we would caution against revising the chapeau of that paragraph, which we believe provides an appropriate and balanced way of portraying those measures. I thank you, Chair.
I thank the distinguished representative of Brazil for commenting on the tax proposals and sharing the views of his delegation. Next on my list is Spain, to be followed by Peru, Austria, the Russian Federation and Greece. I will first pass the floor to the representative of Spain. You have the floor.
Thank you very much, Chair. Please allow me to begin my remarks by making a general comment and then I'll refer to the specific paragraphs we're reviewing this morning. I'd like to reiterate how important Spain and many other delegations feel it is to achieve a consensual document this week. only too aware of the fact that we're at the heart of the discussions on this document and that we are discussing some really crucial paragraphs for the adoption of this document. Throughout the morning we've heard a number of different proposals and new reactions which we consider to be very interesting and they could actually align better with this delegation's view of the subjects we're discussing. But at the same time, and particularly bearing in mind that we really want to achieve a consensual document, Spain's priority is to always consider if we can live with the proposals that you have made as a guide for achieving this consensus rather than trying to achieve what would be a perfect text for us. There's an excellent expression in Spanish, and I apologize to the interpreters in advance, I don't know that if it can be well translated, but it's absolutely relevant for this week and we should always keep it in our minds this week and that expression is perfection is the enemy of good more or less in English so even though there's been some excellent proposals we need to bear in mind this expression because we do need to make sure we achieve a consensus. That's what we need to bear in mind as we go through the document. This is this delegation's priority. Now, I'd like to move on to some specific comments with regard to paragraph 35. We particularly appreciate your inclusion, Chair, of this reference to applicable international law. We think that this is most pertinent because I think that it opens the paragraph up to international law and it is fitting that it does so. We could consider more detailed and specific references to the applicable international law, provided it achieves consensus across all delegations. Now, with regards to paragraph 36, Spain, along with other delegations, would prefer the previous wording, human judgment and control in that order. This is because our opinion, this is one of the cornerstones of the entire exercise. We must underscore the human element. And that human element must be exercised in controlling the system. Nevertheless, we do realise that certain delegations interpret the word control differently and effective control. So we could live with the amendment to control and human judgement provided we include the second part of your proposals chair in which you specifically refer to the control being exercised by human beings either directly or indirectly. By way of conclusion, we interpret this paragraph as a package exactly how you presented it. So Whilst we can live with the amendments, we would prefer the previous wording. We would accept examining Ireland's proposal in more detail whenever it garners broad consensus in the room. Now, with regard to paragraph 37, we haven't got any specific comments there, and we can accept all the proposals that you have put forward. Now, with regard to paragraph 38, the catalogue of measures there listed in the various subparagraphs should not be understood as an exhaustive list. Perhaps this isn't the appropriate place for listing all the possible measures. We believe that there are other possible fora, negotiating fora, in which an exhaustive list of these measures could be discussed. Now, with regard to subparagraph, there are no specific proposals for amendments there, but I would like to indicate that, as far as Spain is concerned, we should follow military terminology and refer only to chain of command, not chain of command and control, because there are no two separate chains. You don't have one chain for command and another chain for control. There's just one chain of command and that has the inherent responsibility of control as well. Now, with regard to subparagraph b, everyone knows that my delegation with others in the room have in previous sessions conveyed the opinion that we shouldn't put a limit on the number of engagements. So, We appreciate your deletion and we are in line with the explanations given by Czechia as to why that should be deleted. With regard to subparagraph c, we have no objections to the drafting you've put forward, Chair, even though we believe that the ideas contained in this subparagraph are already covered by the previous ones. With regard to subparagraph d, whilst there are no new proposals there, we would like to support the relevance of referring to timely deactivation and neutralization. Thank you very much.
I thank the distinguished Ambassador of Spain for commenting on the text and making proposals. Next on my list is Peru, to be followed by Austria, Russian Federation, Greece, Panama. But let me first pass the floor to the distinguished representative of Peru. You have the floor.
Thank you very much, Chair. Peru appreciates the efforts you've made to bring positions closer together on paragraphs 37 and 38, from 35 to 38. Now we support the wording that is taking shape here, particularly with regard to the recognition that control and human element are vital for ensuring compliance with international law. And that it's humans that exercise this control either directly or indirectly. With regard to paragraph 35, whilst our preference would be to keep a reference to international law, in particular IHL, we could also consider the proposal according to which we would refer to the rules of international international law which is applicable to armed conflict. Now this comes from the preamble of the CCW. Now with that in mind, there's another provision in article two of the CCW which could be useful here. And that provision reads as follows. No provision in the current convention nor the protocols and next to it shall be interpreted in a way which would undermine other obligations on the high contracting party, notably with regard to applicable international law to armed conflict. Now with regard to paragraph 36 concerning the human element, whilst Everybody knows that the Peruvian delegation has over the past supported meaningful human control. In the spirit of compromise, we could support the new wording put forward by you, control and human judgment. So this means we consider control as a separate element. Also with The same paragraph 36, we support the deletion of that parting, including the principles and requirements of distinction and so on and so forth, because this would only repeat what's already contained in Article 35, sorry, paragraph 35. I'd also like to recall that Peru in the informal consultations last week, took a special interest in Ireland's proposal because we consider that it is a really good starting point for striving to find convergence across the board. So we're flexible regarding terminology, but this should not be interpreted as giving up on some of the substance. In our view, the text should maintain a clear requirement that human beings retain control over laws and that they should be in a position to determine on the basis of the anticipated effects of the attack that such use will comply with international humanitarian law. Chair, whilst We're flexible on a number of issues. We cannot support the deletion or dilution of the control element, nor any wording which would suggest that the control could be exercised by the system itself instead of by human beings. In short, we'd like to express our support for the text that you have proposed, which I think strikes some really important balances in these paragraphs. And we would particularly like you to consider the comments that we've made on paragraphs 35 and 36. Thank you.
I thank the distinguished representative of Peru for commenting on the text and making proposals. Next on my list is Austria, to be followed by the Russian Federation, but let me first pass the floor to the representative of Austria. You have the floor.
Thank you very much, Mr. Chair. Thank you to you and your team for all the efforts put into putting proposals forward that we can discuss this morning on the screen. With regard to the paragraphs in front of us by now afternoon, paragraphs 35 to 38, we do agree with your assessment that these paragraphs need to be considered together and we do consider them as a package. These paragraphs have been debated for a long time, and we have now heard several proposals on some of these paragraphs, particularly paragraph 36. In a constructive spirit, we will try to work with some of them. On paragraph 35, we were supportive of your text as shown on the screen and would agree that the addition of applicable in front of international law already limits what we are talking about and referred to in this paragraph. On paragraph 36, we have strongly been supportive of language that ties judgment and control to the human actor. The human element needs to be firmly reflected in this paragraph. While looking at your proposal, we already remarked previously that the inversion of the terms are not quite logical to us. A human needs to judge a situation before determining the control to be exercised and may continue to do so in loops. What we cannot accept is any formulation that would insinuate that a one-off pre-intervention authorisation would be sufficient. The aspect of possible changing circumstances and reassessment should therefore be reflected explicitly in this paragraph. We do agree in general with the notion that direct human control does not have to be constant. and we look at the Irish proposal as more favorably and reflecting some of those aspects. They meet some of the concerns that other delegations have expressed and we do see merit at looking at this as a bridge building among different views. We are ready to engage further on this proposal also with regard to the aspect of changing circumstances which was already suggested in an informal by Norway. On para 37, we have previously suggested to strengthen the term entail to require as was referred to by some delegations already this morning. We have also in the informal consultations proposed to add after the operational context, and I quote, the presence of civilians. Even though we consider it as part of the operational context, it would be important to make it explicit here as an essential element. On paragraph 38, we previously made some comments on the subparagraphs, and we have heard in the room this morning that further discussions and more comments on this paragraph and subparagraphs might come at a later stage. And therefore, I will stop my intervention here and stand ready to rejoin the discussion on these paragraphs at a later stage. I thank you.
I thank the distinguished representative of Austria for stating the positions of her delegation and indeed probably we get a further discussion on the powers in 38. Next on my list is the Russian Federation and then perhaps we can also still hear the views of our Greek colleague but let me first pass the floor to the distinguished representative of the Russian Federation. You have the floor.
Thank you, Chairperson. And at the outset I would like to thank you for preparing this new proposal to seek compromise on 35 to 38. We believe that this proposal is a basis for further discussion and we would like to use it as a basis for that discussion because as we see it, this is not so much a step in the right direction towards consensus and mutual recognition of the concerns of all delegations in this room. It is a step in the right direction, we believe. Now on 35. Like many other delegations, we would prefer to keep strictly to the reference to IHL rather than international law as a whole. In the context of these paragraphs on human control, we believe that a reference to international humanitarian law is the more justified reference. On 36. First of all, we have concerns about the second sentence where there are mentions of human control, human judgment and control, and specifically the possibility of carrying that control and judgment out. And our position has not changed on that. We consider that human control and judgment must be carried out against lethal autonomous weapon systems and the specific methods for doing that are at the discretion of the state concerned. Now, given that We welcome the start of the first sentence where it is stated that such control is carried out by human beings directly or indirectly. But the second part of the sentence gives rise to some concerns in the wording that we see here. since we regard this wording as, on the one hand, an attempt to state that human control is not applicable, an applicable criterion for determining whether IHL is being complied with or not, and secondly, an attempt And this is the end of the sentence. We see that as an attempt to consider the forecast consequences of the use of lethal autonomous weapons systems as a mandatory element of carrying out such human judgment and control, including in the context of state application of the standards and principles of IHL. And we consider those attempts to be counterproductive. IHL does not provide as a binding element for carrying out of human judgment or control, at least as it stands now. And furthermore, we do not see a need to try to consider the forecast consequences of the use of lethal autonomous weapons systems as a basic criterion for the conduct of a correct human judgment and control since these forecast consequences can be interpreted incorrectly including in the future and there is a key issue that arises here in connection with this category this legal category who will be deciding to what degree fully correctly Such consequences were forecast at the stage prior to exercise of human control and judgment. Once we've determined that, we can determine who is going to be carrying out the decisions and providing the relevant assessments. Given everything that I have just said, we propose adjusting the second sentence of 36, deleting the middle clause and leaving the following:.
Human beings exercise such control directly or indirectly in accordance with applicable IHL.
In our view that wording will be fully will fully take into account the concerns of all the interested delegations while clearly stating that human control must be carried out directly or indirectly fully in line with and such a wording ensures flexibility for the future should the standards and principles of international humanitarian law be in any way modified. Moving to 37. Overall, this paragraph does not give rise to any objections on our part, but what we don't understand is why at the end we have 'as a whole' added to refer to the weapons system. We find this addition superfluous and would prefer to delete 'as a whole'. On the chapeau of 38 and the new wording, we have no particular objections. Like a number of other delegations, we see these subparagraphs as possible examples of the use of human judgement and control. Within these subparagraphs now, a, we believe that A should be made clearer and we should not be going into the details contained in the second part of it. So we propose placing a full stop after chain of command and control. B, on B, we appreciate the new wording and the deletion of the previous ending. This takes into account the concerns of our delegation and still the beginning of subparagraph b continues to have references to the possible consequences of the use of lethal autonomous weapons systems and we do have concerns about that as I mentioned just before on paragraph 36. And thus, as a compromise, we would prefer to delete the words 'those responsible for the use of laws English'.
Including inter alia by restricting.
And at the end of this subparagraph, just add the words 'can be limited', so it would be as follows:.
'Scale of the operation, types of targets, duration and geographical scope can be limited'.
Moving to sub-paragraph c. Overall, we welcome the amendment that you have made here. However, to be consistent in the terminology, we would prefer instead of "by a human" to say "under human judgement and control" or "under control and human judgment, depending on which of those is in the end selected as preferred by all of the participating states. And on D, we do not have any comments and can accept it as is. I thank you.
Let me thank the distinguished representative of the Russian Federation for his detailed comments and suggestions. And I think we now adjourn for lunch. See you then back in the room at three o'clock. I have some announcements for you to make, but let me first just for a way for you to have an idea what is still going to happen. We have one, two, three, four, five, six, seven, eight, nine, ten, eleven more requests for the floor. So we will continue our discussion on the power of 35, 38. And please, before we break for lunch, let me inform you of the side event organized by Peace Momo and the Friedrich Ebert Stiftung on human control and the crossroads of governance, security exceptions, military AI beyond weapons, and the retreat for meaningful. This side event will focus on the comparison of the EU's and the ROK's respective AI frameworks, examine how human judgment and control can be operationalized, and consider the implications for the CCW GGE on laws process. This side event is open to all from 1:15 to 2:45 in room Concordia 5 in building A of the Palais des Nations. And that's it, and looking forward to see you back in the room at three o'clock.