7th Meeting - 2nd Session Group of Governmental Experts on Lethal Autonomous Weapons Systems 2025
Meetings & Events
Date: 4 September 2025
Language: English
Transcript: https://transcripts.un.org/ar/asset/k13/k13pecnsmg?lang=en
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.

---

Chair [0:00]:

Excellencies, distinguished delegates, I ask you to take your seats again. Good morning, everybody. I hope you had a good night's rest. We are still in box five. And we have at this moment two more requests for the floor from Sweden and the Campaign to Stop Killer Robots. And I see Switzerland, so that's three. Any other delegations who would like to take the floor, please raise your flags. I see there the Republic of Korea also, especially at the back. Good. Let me then now start by giving the floor to the distinguished representative of Sweden. You have the floor.


Sweden [1:08]:

Thank you, Chair, and good morning, everyone. Sweden supports the substance of the Box 5 and appreciate these changes made since our last session. We believe that there could still be some further refinement of formulations, and maybe the text of this box can be shortened without losing its substance. Generally, we prefer a treaty near language in this box, which plays existing obligations of international law in a lost context. Along that line, we support the US proposals to para 2. Thank you.


Chair [1:43]:

I thank the distinguished representative of Sweden. Next on my list is Campaign to Stop Killer Robots. You have the floor.


Campaign to Stop Killer Robots [1:53]:

Thank you, Chair.

We welcome the recognition in box 5 of the need to ensure individual responsibility and accountability. As with our comments on box 4, we welcome the elaboration in box 5 of some of the elements and requirements that can contribute to users' functional understanding of systems, a key part of meaningful human control. and which will therefore promote responsibility and accountability. The only way that meaningful responsibility and accountability can be possible when it comes to increasing autonomy in weapons systems is if there is meaningful human control of our weapons systems. Without such control, accountability for harm from autonomous weapon systems cannot be assigned in a way.

That serves justice to the.

Victims of such systems, nor can it be enacted in a way that will be fair to the system's users.

Only a human user that adequately understands a system and its anticipated effects and applies limits to its use can be reasonably held accountable for the outcomes of using a system.

Robust, clear legal rules to ensure meaningful human control remain the most important tool.

Towards ensuring responsibility and accountability for autonomous weapon systems. We urge states to continue to move forward to enshrine such rules in an international, legally binding instrument.

Thank you, Chair.


Chair [3:40]:

I thank the distinguished representative of Campaign to Stop Killer Robots. Next on my list is Switzerland, to be followed by the University of Utrecht. But first, let me pass the floor to the distinguished representative of Switzerland. You have the floor.


Switzerland [3:56]:

Thank you, Mr. Chair. And I will be short and more or less say what our Swedish colleague just said before us. We do believe that the measures as they are identified in box 5 to ensure accountability and compliance are in good shape already, even though they, in our view, could still be expanded or streamlined a little bit. We do share the view that we do not need to reinvent the wheel. Many instruments foresee explicitly measures with the aim to ensure compliance and accountability, existing instruments, and we think we should inspire ourselves by this existing language, especially if our work should guide us towards a legally binding instrument. For instance, article 14 of amended protocol II of the CCW combines legislative and other measures, mentions imposition of penal sanctions for certain acts, instructions, procedures and training, as well as procedures for consultations and co-operation to resolve problems that may arise with regard to interpretation or implementation. Also, the Geneva Conventions, as well as their Additional Protocol I, serves as another example for instruments containing provisions for the execution of the Convention on dissemination, the enactment of national legislation to provide penal sanctions, or on responsibility. Thank you, Mr. Chair.


Chair [5:52]:

I thank the distinguished representative of Switzerland for stating the position of his country. Next on my list is the University of Utrecht, followed by the Republic of Korea, the European Union and Hungary. But first, the distinguished representative of the University of Utrecht. You have the floor.


Utrecht University · APILS [6:10]:

I thank you, Chair. I make this joint intervention on behalf of the Asia Pacific Institute for Law and Security, APILS, and on behalf of my own delegation, Utrecht University. We note that the current version of the box five text is much more consistent in the use of the terms responsibility and accountability than previous versions. We commend you, Chair, and your team for ensuring this alignment. The substance of this box has received wide support from various delegations, though, as the Swedish delegation also saw, we see some room for making the text more concise and streamlined. Chair, our comments dovetail on some interventions yesterday proposing the replacement of the word accountability with responsibility. And in our view, that change would unnecessarily narrow the mechanisms for ensuring compliance with international law that this box aims to achieve. Responsibility refers specifically to forms of accountability based in law, such as through state responsibility or individual criminal responsibility. Accountability refers more broadly to processes in which an accountability holder holds a power wielder to a set of agreed standards. In short, responsibility is just one mechanism for accountability. Accountability in the context of laws refers to the requirement for actors to explain or justify their actions or omissions. with potential political, legal, social, or other consequences. It does not necessarily lead to or depend on legal proceedings to incentivize compliance. Not every violation of IHL entails individual responsibility under international law, so other avenues of accountability may, in our view, be appropriate in certain circumstances. Limiting accountability to mechanisms of responsibility under international law misses an opportunity to cast a wider net to shape state conduct in relation to laws throughout the rolling text. Chair, with this in mind, we recommend the following changes to the text, and we offer these suggestions in the spirit of streamlining and clarifying the intended meaning of the text. We will, of course, submit these proposed changes in writing to you for your team and for the review. We think that paragraph one is well written and does not need further changes. We suggest simplifying paragraph two. First, as suggested by several delegations yesterday, it would be appropriate to delete the word "must." Also, the second sentence of the paragraph creates confusion. It's not clear what it means for accountability and responsibility to be, quote, "considered across the life cycle of a loss." Rather, states parties to armed conflict remain responsible for all of their decisions across the life cycle. Therefore, elements of the second sentence could be incorporated into the first. We propose the following revised language for paragraph two. States, parties to armed conflict, And individuals remain responsible and accountable in accordance with applicable international law for decisions with regard to laws throughout their life cycle, because responsibility and accountability cannot be transferred to machines. We see significant duplication in paragraphs three and four. In effect, paragraph four expresses in more detail the ideas captured in paragraph three. Also, paragraph three makes an unclear reference to the means or methods of warfare involving the use of laws. So we recommend merging existing paragraphs three and four to remove this duplicity and clarify intended meaning within the text. We propose the following wording. States must ensure accountability and responsibility at all stages of the life cycle of laws. including by operating such systems within a responsible chain of human command and control to ensure compliance with applicable international law, including IHL. We agree with proposals made yesterday that paragraph five could be clarified and streamlined. But first, the chapeau contains the operative word provide, which does not fit well with the substance of subparagraph A. States will often not be designing human machine interfaces. so they would not be in a position to provide interfaces meeting certain requirements. Rather, they should ensure that systems developed by industry have such interfaces. Furthermore, we see an overlap between subparagraphs and, and we suggest language from subparagraph to be merged into subparagraph. We also suggest referring to regular training, not just training in subparagraph. We propose the following text. To promote human responsibility and accountability in the use of laws and to mitigate the risk of unintended engagements, states should: A, ensure that laws have readily understandable human-machine interfaces and controls. B, issue guidance and instructions, for example, policies, doctrine, procedures, and rules of engagement. consistent with applicable IHL for personnel regarding the proper use of laws. And C, provide regular training of personnel to understand such guidance and the capabilities and limitations of the laws on autonomous functions and the anticipated circumstances of its use. We think that paragraph five is well written and does not need further changes. And finally, in paragraph six, we suggest adding the words or control at the end of the sentence. This change in language reverts back to previous versions of the rolling text and reflects formulation in existing international law, including the amended Protocol II of the CCW. We thank you, Chair.


Chair [12:05]:

I thank the distinguished representative of the University of Utrecht for a joint statement. Next on my list is the Republic of Korea, to be followed by the European Union. Let me pass the floor now to the distinguished representative of the Republic of Korea. You have the floor.


Republic of Korea [12:21]:

Thank you, Mr. Chair. At the outset, my delegation would like to warmly welcome Czechia for endorsing the draft articles, a document that the Republic of Korea, along with other states, has supported, and its circulation as a working paper to renew attention to this important text. We also welcome the many elements of box 5 are drawn from this working paper, affirming that the humans will remain responsible and accountable. And we express our general support for box 5, agreeing with the chair that this box is in a good shape. We also support further streamlining this box as suggested by the US delegation yesterday and others today. Going back to yesterday, I would like to return to the comments made by the US delegation regarding box 4. While we need to carefully study the specific textural edits proposed by the distinguished delegates of United States, we believe that the grouping measures applicable to the development phase is a good idea worth giving a consideration. We agree that certain principles should apply throughout the entire lifecycle of laws. However, for the sake of drafting a text that is reader-friendly, logical, and coherent, one that will appeal to the broader public and future diplomats, we should consider reconstructing, restructuring this portion of box five, box four, sorry, bringing elements related to the development phase from different parts of the text. Thank you.


Chair [13:54]:

I thank the distinguished representative of the Republic of Korea for making proposals and suggestions. Next on my list is the European Union, to be followed by Hungary, and then I have Australia, Venezuela, China, and Spain. But let me first pass the floor to the distinguished representative of the European Union. Madam, you have the floor.


EU [14:17]:

Thank you, Mr. Chair. Since it is the first time that I take the floor, I would like to thank for your leadership of the group, as well as for yours and your team's efforts and work on the rolling text, also during the intersessional period. At previous GG&L sessions, we presented examples of risk mitigation measures, including readily understandable human-machine interfaces and controls, comprehensive and systematic training of personnel, establishing doctrine and procedures, and circumscribing weapons use through rules of engagement. We are glad to see that they have been reflected in the current version of the text. With regard to paragraph 6 and 7 of box 5, the EU emphasizes that to ensure and maintain human responsibility and accountability at all times and across the entire life cycle of the weapon system, In accordance with States' international obligations and commitments, States may consider adopting and updating appropriate national measures, which could include the exchange of information or data among the competent national authorities, investigating incidents that may involve violation of international humanitarian law, perpetrated by their armed forces and nationals or on their territory, including through such weapons systems, as well as to undertake appropriate actions in response to such investigations. I thank you, Mr. Chair.


Chair [16:22]:

Let me thank the distinguished representative of the European Union for the points brought Forward, next on my list is Hungary to be followed by Australia. But let me first pass the floor to distinguished representative of Hungary. You have the floor.


Hungary [16:38]:

Thank you, Mr. Chair.

Good morning, everyone.

Hungary values the contents of the box five of this rolling text.

The measures are in good shape in order to establish and maintain the compliance to international humanitarian law.

We propose the following text for para 4 for the implementation of safeguards in order to make it.

Clearer that states can have dual obligations regarding lethal autonomous weapons system.

First, to ensure the weapon is capable of being used lawfully, technical compliance, and second, to ensure it is used lawfully under responsible human oversight, operational compliance.

The text is the following.

For paragraph four.

States must ensure that lethal autonomous weapons systems are capable of.

Being used in strict compliance with international law throughout their life cycle, including after any modifications.

Furthermore, states must ensure full accountability for their use through a responsible chain of human command and control. Thank you, Mr. Chair.


Chair [17:46]:

I thank the distinguished representative of Hungary for making proposals with regard to the text. Next on my list is Australia, to be followed by Venezuela and China. But first, the distinguished representative of Australia. Madam, you have the floor.


Australia [18:03]:

Good morning, Chair. We appreciate your ongoing work to refine Box 5. We welcome the new paragraph 5, which represents a practical improvement to aid states in promoting responsibility and accountability for the use of laws. We have a number of comments across box 5 to provide clarity to the text and ensure it remains faithful to international law. In our view, there is currently a lack of consistency within the language in box 5 around accountability. In particular, the text blurs the concepts of state responsibility and individual accountability. It is important that box 5 clearly sets out how states, parties to armed conflict, and individuals can be held accountable for violations of IHL. IHL imposes international obligations on states. Therefore, a state is obliged to ensure compliance with IHL. It is its responsibility. IHL does not impose obligations on individuals. Rather, under international law, States are required to ensure individuals who commit IHL violations are held to account through their own internal mechanisms. On paragraph 1, we agree with the suggestion made by the United States and other delegations that paragraph 1 be moved to box 2, as it reflects a useful general principle agreed by this GGE. On paragraph 2, We believe this paragraph should be confined to obligations on states and parties to armed conflict only. It is unclear what the reference to individuals in paragraph 2 is seeking to achieve. We suggest individual accountability should be addressed separately in paragraph 3. Additionally, we suggest this paragraph should be limited to decisions with regard to the use of laws. IHL regulates the use of laws. it is not clear how there could be accountability for any decision relating to laws. Turning now to paragraph 3, which concerns individual accountability, we suggest removing the reference to means and methods of warfare, as it is unclear what this is intended to capture in this context, and there seems no reason to limit individual accountability in this way. Further, we suggest that it would be appropriate to refer to obligations under international law rather than IHL, given individuals are held accountable for violations of IHL through international criminal law. On paragraph 4, we support inclusion of the phrase "responsible chain of human command and control." The edits seem to have resolved previous issues raised with this paragraph. and more closely adheres to consensus language in guiding principle D. In line with our proposed amendments to paragraphs two and three, we would propose similar minor edits to paragraphs four and five to ensure the distinction between state responsibility and individual criminal accountability remains clear. In respect of paragraph six, we suggest removing the second sentence. The first sentence is not prescriptive as to where in the life cycle an incident may occur, and so we do not think the second sentence adds anything material to the first. We would also suggest "and appropriate action in response to incidents" instead of the term "redressal". Finally, we suggest updating phrases within the rolling text which are no longer used consistently. For example, "weapon system and such systems" can be replaced with "laws", and "entire life cycle" can just be "life cycle". We will submit our textual edits on this box to you and your team, Chair. In our view, these suggestions would go a long way to improving box 5, ensuring that states and individuals remain at all times responsible for adhering to their obligations under applicable international law, including IHL. I thank you.


Chair [22:21]:

I thank the distinguished representative of Australia for making proposals with regard to the text in box 5, and we would appreciate to receive them also in writing. Thank you for that in advance. Next on my list is Venezuela, to be followed by China and Spain. But let me first pass the floor to the distinguished representative of Venezuela. So you have the floor.


Venezuela (Bolivarian Republic of) [22:48]:

Thank you very much indeed, Chair. We are grateful to be able to speak to the plenary. Our delegation is thankful for the presentation of this part of the document, which embodies important elements for our country. We think it is necessary to include some of the elements that strengthen our proposal and would like to make the following comments. Autonomous lethal weapon systems raise a number of ethical, legal, moral and technical issues. the best technologies in the arms area in general have only been used to increase their killing potential. Because of the lack of legally binding instruments, there is no responsibility or accountability, which is why we welcome the proposal contained in paragraph 5. With regards to a, b, c and d thereof, we should take into account the fact that all weapons, including those that have autonomous functions, must remain under the control and the direct supervision of human beings at all points. To delegate the decision on life and death is not only not ethical, but it also raises serious challenges for adherence to international humanitarian law and contravenes the consistency mentioned in b on the applicability that is, the policies, doctrine, and procedures. We should also consider in 6 and 7 that we must take into account the potential challenges and indiscriminate effects which laws and their use might have in an armed conflict, stemming from, inter alia, one, the inherent risk that these weapons can have in these effects, as if they're not been designed to identify and attack a single objective. Number two, the inability of these weapons to fulfill IHL in proper judgment with regards to distinction, proportionality, and precaution, these main. Three, those who decide and plan an attack, that responsibility cannot be transferred to a machine. Four, the need to ensure responsibility of states, accountability for international illicit acts for the use of these weapons. Thank you very much, Mr. President.


Chair [25:24]:

I thank the distinguished representative of Venezuela for stating the positions of his delegation. Next on my list is China, to be followed by Spain and Bangladesh. But let me first pass the floor to the distinguished representative of China. Madam, you have the floor.


China [25:42]:

Though we understand fully the request from you of not introducing so many new elements or new language into the current draft, China would like to propose to add one more sentence or paragraph reaffirming para J of 2019 guidelines, which is on the peaceful use of autonomous energy technology. will be, I can read the sentence we propose is as follows. Measures taken within the context of the CCW should not hamper progress in our access to peaceful uses of intelligent autonomous technologies, which is exactly the guiding line J from the 2019 report. And we are quite flexible with the place where the sentence goes. It can either at the end of the box or at the end of the whole draft. Thank you so much, Chair.


Chair [26:52]:

I thank the distinguished representative of China for her intervention, making proposals. Yeah, um, yeah, the, the placement would be a little bit difficult in, in power, um, uh, responsibility and accountability, I would say, but on the other hand, we, um, uh, We have a text and we have earlier said also with regard that we now not embark on discussions, moving text around boxes, because that has to be revisited anyhow. So, thank you for making this proposal. Next on my list is Spain, to be followed by Bangladesh and Malawi. But first, the distinguished representative of Spain, you have the floor.


Spain · Ambassador [27:36]:

Thank you very much, Mr. Chairman. briefly, first of all, to support this text in box B.5, based on the references made to the accountability for individual states, and those references we support. Two more detailed comments on 5 and 6, which also apply to 4. Sorry, 4, not 5. Some delegations we have noted have suggested to merge or blend 4 with other paragraphs. We would like to highlight the need to keep this paragraph as a standalone one. whether ideas from other paragraphs are added or not. So, to preserve in 4 the reference to the need for these systems, there should be a responsible human chain of command. Some proposals for mergers with other paragraphs do not embody that reference, and we think we absolutely must keep that reference in 4 as we read at the end. in the last line. With regards to paragraph 6 now, we've heard some delegations propose that we should get rid of 6. We believe that a paragraph that refers to the federated states must ensure effective mechanisms for investigation, reporting and redress of incidents that may involve the use of laws is very important. Not only is it important in itself, but it is important because it is consistent with and gives body to the principles of responsibility and accountability which prevail throughout international law and especially in international humanitarian law and in our very own domestic legal machinery in Spain. So my comment on six also applies to paragraph 7. Generally speaking, Mr. Chairman, we are grateful for paragraph 5 and we like the way it has been drafted. Thank you very much.


Chair [30:18]:

I thank the Ambassador of Spain for putting forward the positions of her delegation. Next on my list is Bangladesh, to be followed by Malawi, the United Kingdom and Israel. Let me first pass the floor to the distinguished representative of Bangladesh. Sir, you have the floor.


Bangladesh [30:35]:

Mr. Chair, thank you for giving me the floor. I am presenting additional arguments on the issue of individual accountability involving the use of laws. as this has surfaced again in para 2 and 3 of box 5 of the rolling text. Yesterday, the distinguished US delegate was referring to IHL not being devoid of accountability provisions. While echoing that viewpoint, my delegation understands it as a reference to the entirety of international criminal justice in warfare. We believe that it will not be prudent to conclude an instrument on regulation of laws without taking this perspective into consideration. This specialized area within international law ensures that war is not a lawless space and is therefore key to ensuring justice and accountability in warfare. In so doing, IHL and ICL work in tandem and complementarily. IHL sets the rules of armed conflict, and ICL enforces accountability for breaking those rules. In fact, the function for a single mission, deterring grave violations of rules of armed conflicts. Despite their different functions, they are indeed mutually reinforcing. Weakening ICL will lead us to the point of time when IHL suffered from the lack of its teeth, a robust accountability mechanism. Mr. Chair, LOWS poses a serious threat to justice and accountability in warfare. International criminal justice deals with one of the key legal questions of attribution of responsibility. Delegating lethal decisions to machines breaks the link between human decision-making and accountability, which is essential under both IHL and ICL. Attribution of responsibility for grave violations of IHL then becomes a daunting challenge under ICL. Responsibility cannot be outsourced to algorithms. International criminal justice has recognized criminal responsibility of humans only. Machines are devoid of legal personhood, criminal intent, and moral judgment. Machines do not stand trials, nor do they go to prison. People do. And if no one can be held accountable for unlawful killings, we are not just breaking the law, we are rewriting it in a way that permits impunity. This is such an accountability gap again being created by laws, as rightly pointed out by the distinguished delegate of Austria yesterday. Mr. Chair, international community has walked a long path in closing this accountability gap in international law. They have broken this culture of impunity by holding individuals criminally responsible for international crimes, including war crimes. On the other hand, we repeatedly stress that in terms of accountability, LAWS is truly unique compared to other weapons systems. LAWS, by its very autonomous nature, has stood in our path by exempting individuals from criminal responsibility. LAWS have seriously challenged key principles of accountability, responsibility and attribution that underpin the core of international criminal justice. The ruling text should therefore not overlook this diametrically opposite relationship between international criminal justice and laws. If overlooked, this will significantly weaken the principle of individual criminal responsibility by returning and sustaining that culture of impunity. Without clear mechanisms for assigning legal liability, laws risks not only eroding the legal norms that govern armed conflicts, but also those that ensure justice against other international crimes. Mr. Chair, the regulation of laws therefore warrants our renewed commitment to the principle of individual criminal responsibility. My delegation reiterates our position in this regard with a slight change in the text we proposed earlier. Humans, not machines, bear individual criminal responsibility for grave violations of IHL. I repeat. "Humans, not machines, bear individual criminal responsibility for grave violations of IHL." This can be placed in either box 2 or in box 5, para 2, after the phrase "since responsibility and accountability cannot be transferred to machines." Mr. Chair, my delegation believes that its incorporation in the text is possible. Firstly, the rolling text has reference to international law beyond IHL. Secondly, CCW has reference to the principles of international law beyond IHL. Thirdly, delegations numerously talk about responsibility and accountability in the regulation of laws. Even if the delegations are still hesitant about incorporation of individual criminal responsibility as an IHL principle, let me assure you that it is also recognized as an IHL principle. Geneva Convention Article 146 and Additional Protocol I, Article 85, elaborated on this. Rule 151 of ICRC's Compilation of Customary International Humanitarian Law, Volume 1, clearly affirms that individual criminal responsibility, specifically for war crimes, is a customary rule not limited to treaty contexts and applicable across both IAC and NAC. The verdicts of the international criminal tribunals have also upheld it. My delegation urges this august body to reflect this vital principle of international criminal justice in the rolling text. As laws directly challenge it, the best way to counter it is to renew our commitment to this principle. Thank you, Mr. Chair.


Chair [36:22]:

Let me thank the distinguished representative of Bangladesh for highlighting the issue of criminal responsibility. Next on my list is Malawi, to be followed by the United Kingdom, Israel, and now also France and Panama. But let me first pass the floor to the distinguished representative of Malawi. You have the floor.


Malawi [36:47]:

Thank you, Chair Persson. Chair, I would like to beg your indulgence to comment on box four. My apologies, it's a small mission problem, so we have to be in different places. So on box four, we'd just like to support a few propositions that were made yesterday by other delegations. We support the replacement of deployment with employment in the chapeau. On paragraph one, we support the delegation of Mexico for the legal reviews to be conducted periodically. On the same paragraph, we commend the inclusion of voluntary exchange of relevant practices, which can also be conducted in a broader context in form of international cooperation, as it is done in other instruments, because this will boost confidence and bring convergence of standards. On paragraph 5, We appreciate the inclusion of rigorous testing and evaluation, including simulation of its use in operational environments. However, we would like to propose an addition after the phrase "environment" that we put "taking into consideration different geographical contexts." On paragraph 6, we would like to support delegations that propose for the explicit listing of forms of biasness, in particular in terms of race and gender. Now, commenting on box 5, we believe that legal and moral responsibility must always rest with humans and States and not machines. Therefore, we support the current text. We believe that it has captured the whole essence of of having States being responsible and accountable. Thank you.


Chair [39:00]:

I thank the distinguished representative of Malawi for the comments shared on both boxes four and five, and you are totally free to do so, even much larger delegations do so, so why should you not do it? Next on my list is the United Kingdom, to be followed by Israel, but first the UK. You have the floor, sir.


United Kingdom of Great Britain and Northern Ireland [39:20]:

Thank you, Mr. Chair. The UK would like to note that, both this morning and throughout discussions that we have had in the course of this mandate and even before, the role of accountable actors has been central to the points that we are seeking to conclude within this room. It has been so axiomatic that we appear to have reached consensus repeatedly on this point in the reports of this group and in the guiding principles. It highlights the criticality of these issues of accountability. The UK's position is that the discussion of the role of the human cannot be properly understood without the understanding of those who are accountable. The UK strongly agrees with the suggestions that we've heard from Bangladesh and others that the points about accountability not being held by machines needing to be made in the structural positions in this text. So therefore, reflecting points that we see in paragraphs one and two, should be moved higher into the text, into box 2, as it's necessary to address the logical nexus between the understanding of what systems we are talking about and what the role of the human is. The discussions that we've had about context-appropriate human judgement and control have been somewhat contentious. But overall, the issue around the necessary level and form of human engagement with the system is framed by the fact that the human is the accountable actor and acting in the context of use and in the circumstances. We do not have much by way of textual amendments at this stage to this box, beyond what we have heard in this room. We note that the reference in paragraph 4 in relation to the operation of systems within a responsible chain of human command and control could be further clarified by the inclusion of "those governed by" to read States must ensure accountability across the entire life cycle of laws to ensure compliance with applicable international law, including through the operation of such systems by those governed by a responsible chain of human command and control. Thank you, Mr Chair.


Chair [41:49]:

Let me thank the distinguished representative of the UK for making suggestions and also with regard to the structure that we indeed have also heard proposals from Bangladesh and for your suggestion to have additional text in paragraph four. Next on my list is Israel, to be followed by France, Panama, Ukraine. But first, distinguished representative of Israel, you have the floor.


Israel [42:20]:

Thank you, Chair, and good morning. Regarding the third paragraph in box five, there are some duplications in our view between paragraph three and paragraph four. So in order to streamline the text, And since paragraph four is based on principle D of the 2019 Guiding Principles, we believe it would be preferable to delete paragraph three. Regarding the sixth paragraph, the paragraph addresses a highly important subject. However, it exceeds the scope of work of this group and introduces new terms that have not been the subject of thorough discussions. Therefore, we suggest deleting this paragraph at this time. However, if the paragraph is retained in the text, we would suggest replacing the word "must" with the word "should," as it would better reflect the practical nature of this text. Regarding the seventh paragraph, there seems to be some redundancy between this paragraph and paragraphs 3 and 4. So in order to streamline the text, we suggest considering to delete this paragraph. However, if this paragraph remains, we propose the addition of the words "such as" after "States must take all appropriate steps," since legislative measures will not always be considered appropriate. Furthermore, we believe that it is important to clarify that we are talking about IHL obligations that are applicable to laws and don't just relate to it. Therefore, we would suggest replacing the word "relating" with the words obligations applicable, so that the text would read, "States must take all appropriate steps, such as legislative and other measures, to prevent and suppress violations of IHL obligations applicable to laws by persons or on territory under their jurisdiction." I thank you.


Chair [44:26]:

I thank the distinguished representative of Israel for the suggestions made. Next on my list is France, to be followed by Panama. Let me first pass the floor now to the distinguished representative of France. Please have the floor.


France [44:42]:

Thank you very much, Chairman. I'll be very brief. Our delegation would just like to express our support for Box 5 and its paragraphs. as presented, which we feel accurately reflects the discussions we have had on the issue of responsibility and also existing obligations. This morning, we have heard a number of comments and proposals for amendments, which we need to look into. Thank you.


Chair [45:14]:

I thank the distinguished representative of France for stating the position of her delegation. Next on my list is Panama, to be followed by Ukraine and Belarus. But let me first pass the floor to the distinguished representative of Panama. Madame, you have the floor.


Panama [45:29]:

Thank you very much, Chair. We didn't want to take the floor to start with because for this box, We were actually quite happy with the wording that had been presented, but we would like to respond now because the issues at stake here are extremely important. and we believe that one of the concerns that exists in the context of discussions surrounding laws is how to avoid impunity therefore it is very important to guarantee not only state responsibility but also individual responsibility and accountability and so that's why we fully support the extraordinary explanation that was given by the distinguished representative of Bangladesh this morning regarding just how important it is to reflect those dimensions in the text. Therefore, we would support keeping all of the paragraphs as they were drafted, and especially as regards paragraph 4. would want to keep paragraph 5, 6 and 7 as well, including the explanations made by the distinguished delegation of Spain regarding paragraph 6 and 7. We do welcome the inclusion of paragraph 5 and its subparagraphs because we feel that they provide practical and useful guidance for implementation We would even be ready to replace the word should by must in the chapeau, although we understand that the intention is not to give an impression of the text being prescriptive. We would like to focus on the importance of ensuring responsibility and accountability, so we feel that we could make a change for the better. But we would like to put down on record our support for keeping all of the paragraphs in this box. Thank you.


Chair [47:49]:

I thank the distinguished representative of Panama for stating the position of her delegation and making suggestions. Next on my list is the distinguished representative of Ukraine. So you have the floor.


Ukraine [48:02]:

Thank you, Chair. The delegation of Ukraine would like to reiterate its call expressed yesterday to remain cautious in establishing any rules of responsibility of states or individuals within the future instruments on the laws, and would like to comment on certain specific aspects of the existing crawling text, box five. First and foremost, we find rational the positions of the distinguished delegates of Sweden and Switzerland. who mentioned the existing accountability and responsibility mechanisms within which our delegation would like to prefer to stay. Our delegation does deem it possible to clarify how some of those mechanisms apply to the laws, but at that, we must remain strictly within the boundaries of the existing mechanisms. Paragraph 1, to our opinion, serves as an incarnation of the general understanding achieved by this group that artificial intelligence cannot replace humans, including in matters of responsibility. One of the elements of this responsibility is to clarify that only humans can exercise ethical and legal judgment, which a machine is incapable of. Therefore, we treat with understanding the proposals made by the United States, Australia, and the United Kingdom, and other delegations, to move this paragraph to box 1, is reflective of a very general principle. To our delegation, paragraph 2 looks like a further explanation of the meaning of paragraph 1, as rightly pointed out by the representative of the University of Utrecht. And it is precisely here where we step on an uncertain ground. This paragraph refers to applicable international law for decisions with regard to the loss. but what's the applicable law. For certain, manuals and the responsibility of states have reached customary status, as has been proclaimed by the ICJ with regards to some particular norms in some particular cases, and tacitly recognized by the General Assembly in 2001 when it recommended the articles and responsibility of states for international wrongful acts for the use in the relations between the states. However, there appears to be no general agreement yet as to the list of such norms on state responsibility. It seems to our delegation that this paragraph hints to the rules relating to the attribution of a conduct to a state as is described in Articles 4 and 11 of the Articles on Responsibility of State developed by the International Law Commission. But further, the paragraph addresses certain decisions with regards to the laws. Does this mean human decisions with regard to the laws or the quasi-decisions of the laws itself? We find the first presumption risky and the second one absolutely impossible for the purposes of establishing a state responsibility. Furthermore, our delegation has doubts that all states would eagerly adopt for themselves an obligation either not yet fully agreed upon within the context of the development of the RCIA or absolutely impossible if we are speaking about placing responsibility on the robots, which contradicts also paragraph one of this box. We therefore propose to refrain from any wording that might hint towards responsibility of states for internationally wrongful acts. This is not because our delegation does not recognize the customary or emerging norms on the responsibility of states. Such norms remain applicable. It would strongly advocate or avoidance of any legal risks which might undermine the recognition of our potential instrument. Speaking of paragraph 3, we find it reflective of the overall duty of states to ensure respect for IHL. We take note of the position expressed by the distinguished representative of the United Arab Emirates to consider adding preference to common article one of the Geneva Conventions. We believe that it could be specified that the duty to ensure respect for IHL is mentioned in common article one, entails an obligation to set the grounds for individual accountability for the lethal autonomous weapons systems. Coming to paragraph four, our delegation believes it to be a continuation of the same idea. Again, we note that this issue has already been observed by the distinguished representative of the University of Utrecht. Adding preference to responsible chain of command and control is tempting indeed, at least from the point of view of the necessity of adding reference to human command and control. On the other hand, we once again call upon the distinguished experts to be cautious, as this wording approaches very closely to the wording of international criminal law. Ukraine supports the expansion of the scope of international criminal law And Ukraine has proven it a year ago by joining their own statute, something that no state has ever done before in the course of an ongoing armed conflict. Yet Ukraine understands that some states do not find it possible for themselves to accept such commitments at this moment. Therefore, it appears to us to be unwise to introduce such commitments within the instrument on the laws. We do not entirely oppose the existing wording, but just want to bring the attention of the delegates to the issue which may potentially be problematic. As for paragraph 5, we find it notable, yet probably belonging to the previous box 4, as it encompasses duties which look like specific precautions. Here, we fully understand the position of the distinguished delegate of the United States concerning getting automation bias to the support of this paragraph. the opinion of our delegation, mitigation of automation bias is a part of precautions. We believe that the paragraphs of paragraph 4 confirm that this paragraph in its entirety belongs to box 4. For example, Ukraine has already underlined the importance of education and training when dealing with paragraph 8 of box 4. Repeating this in a box concerning responsibility may be somehow confusing for the final structure of the document. We therefore propose reconsideration of paragraph 5 in the light of similar provisions in other boxes of the existing rolling text. Our delegation has no specific comments as for paragraphs 6 and 7. However, we would like to specifically underscore our position to certain proposals voiced yesterday. as to the deletion of these paragraphs for alleged reasons of their intervention into internal affairs of states. In our opinion, all states have mechanisms for investigation of violation of law on their respective territories. And if they are not intentionally acting in bad faith, they are themselves capable of preventing and suppressing violations of law on their territory and in territory under their control, or even those committed by their nationals on the territory of other states. Therefore, these paragraphs will, to our opinion, serve as a useful reminder about such duty of every state. With this, let me please conclude. Thank you, Chair.


Chair [55:47]:

I thank the distinguished representative of Ukraine for stating the position of his delegation and making propositions with regard to the text in box 5. Next on my list is Belarus, and I have two more requests for the floor from Belgium and Croatia. Let me first pass the floor to the distinguished representative of Belarus. Sir, you have the floor.


Belarus [56:10]:

Um, which was question?

Thank you very much, Chairman. Our delegation hadn't intended to take the floor. But given how important the discussion is and the points raised by Bangladesh and other delegations about responsibility and accountability. Also, in light of the fact that this issue refers to our proposal relating to paragraph 5 in section 2, we would like to note that we are prepared to discuss this and are flexible regarding the formulation. We just think that it is an important matter, and we also think that it is a complex one. We would like to make the following comments. I don't think that people disagree with the fact that states bear absolute responsibility for the use of laws. But it could be very complicated when it comes to the issue of individual responsibility and accountability, especially criminal accountability. Why? Well, because laws can, different laws can have different degrees of automation and therefore different systems may require different degrees of responsibility of an individual in similar situations. So the systems themselves might be very different. And states, in the framework of their own criminal legislation, know the specifications of their laws, and they are well placed to identify the responsibility and accountability of individuals. But if we're speaking about the international level, then I think we may encounter difficulties because these systems, lethal autonomous weapons systems, may be classified Some of these systems may be well known as a type of weapon, and in that case, international rules regarding responsibility can be applied. But I don't think we can expect that in the near future, all lethal autonomous weapon systems will be equally known in terms of how they are built and composed, and Secondly, they're unlikely to be kept secret, but it is not excluded that some mechanisms and systems may be available on the market and their characteristics may be known, whereas others won't be. So the responsibility will largely depend on that aspect of the relevant system. But there is not a single standard that could apply because all of these systems can be significantly different one from the other. And so that is what I wanted to point out for the discussion. And I'd like to underscore the issue of absolute responsibility of a state is not called into question. That is customary in international law. States bear responsibility for these weapons and similar ones, but I think when it comes to individual responsibility, we have to be very careful in how we approach it, both at the national and at the international level. Thank you.


Chair [1:00:02]:

I thank the distinguished representative of Belarus. Next on my list is Belgium, to be followed by Croatia. and any other delegations who would like to take the floor. I see Canada. Please raise your black. But let me first pass the floor to the distinguished representative of Belgium. Sir, you have the floor.


Belgium [1:00:21]:

Mr. Chairperson, box 5 contains important key elements in terms of responsibility and accountability with regards to laws. We concur with what our Australian colleague stated previously. arguing that box 5 would benefit from being revised to better distinguish between the different types of accountabilities at stake, namely states accountability and individual criminal accountability.

Belgium is prepared to consider the specific amendments proposed by Australia in this regard.

In particular in relation to paragraphs 2 and 4. Furthermore, we would like to underline the need to consider throughout box 5 accountability under the applicable international legal framework in general, and not solely limited to IHL violations. I thank you, Chair.


Chair [1:01:16]:

I thank the distinguished representative of Belgium for referring also to remarks made earlier by other delegations. Uh, next on my list is, uh, Croatia to be followed by Canada and Colombia, um, but first now the floor to the distinguished representative of Croatia, um, yes, Croatia, you have the floor.


Croatia [1:01:37]:

Thank you, Mr. Chair, uh, and thank, uh, to your team for the excellent management of this meeting, it is very demanding, uh, duty. When we speak about box 5, we think that it generally defines responsibility when making decisions and use of laws. We believe that in accordance with the existing legal framework, responsibility is individual and should be regulated as such. In this regard, it is necessary to further discuss the role of the state of which the individuals are the citizens. In this context, we need to be clear about the role of the state as the orderer for the actions of members of its armed forces. Thank you.


Chair [1:02:25]:

I thank the distinguished representative of Croatia. Next on my list is Canada, to be followed by Colombia. But let me first pass the floor to the distinguished representative of Canada.


Canada [1:02:41]:

Thank you, Chair, and good morning, colleagues. My delegation offers the following comments on box number five. Canada would like to reiterate the utmost importance of human accountability in human-machine interaction. Its inclusion fosters greater caution and awareness, and in turn, the reduction of the risk of automation bias. My delegation agrees with Australia's statement on para. 1 to move this paragraph to box 2. IHL imposes obligations on states and parties to an armed conflict and not individuals. States and parties to an armed conflict are expected to hold individuals subject to their jurisdiction accountable for violation of IHL. Paragraph 2, we recommend the deletion of the words and individuals must at all times remain, given that decisions related to laws are that of states. Therefore, the para would read, states and parties to armed conflict are responsible and accountable in accordance with international law for decisions with regards to laws since responsibility and accountability cannot be transferred to machines. They will be responsible and accountable for their decisions during the life cycle of the weapon system. Paragraph 3. We believe that this paragraph can be simplified by removing the words "means and methods of warfare involving." Therefore, the para would read as follows, states must ensure individual accountability for the use of its laws in accordance with their obligations under IHL. Paragraph four, we recommend removing human from the phrase responsible chain of human command and control to ensure the phrase is consistent with other uses of the phrase in the rolling text. Paragraph 5, the nature of the sub-paragraphs is to assist in the responsible use of laws. Therefore, we recommend that the chapeau be reworded as follows: "To promote the responsible use of laws and to mitigate the risk of unintended engagements, States should provide..." and then the rest. We can simplify some of the sub-paragraphs by combining para b and d. The combined sub-paragraphs would read as follows: guidance and direction, policies, doctrine, procedures, rules of engagement, directives, and orders, consistent with applicable IHL for personnel regarding the proper use of laws. Paragraph 6. While we agree in principle with paragraph six, we offer the following wording. States have a responsibility to establish effective domestic mechanisms for investigating, reporting, and addressing incidents that may involve violation of IHL. These include, where relevant, incidents arising from the development and use of laws. In para seven, while we agree in principle with para seven, We recommend the following wording: States are required to take legal and other measures to prevent and suppress violation of IHL, including those related to the use of laws by persons within their territory or subject to their jurisdiction. I thank you, Chair.


Chair [1:06:20]:

I thank the distinguished representative of Canada for his comments and proposals made. Next on my list is the distinguished representative of Colombia. You have the floor.


Colombia [1:06:36]:

Gracias.

Thank you. Colombia would like to thank the Chair and his team for having presented a proposal that brings together essential elements for ensuring accountability in the context of the use of laws. We would like to offer a few comments regarding box 5. We would like to point out that in the context of responsibility, we welcome the inclusion not just of states, but also of parties to armed conflicts and individuals. This vision broadens responsibility to include the idea that responsibility applies not only when laws are under the effective control of a state, but also when they are used by non-state actors, military companies, and private security companies when they may be involved in international humanitarian law violations. Therefore, we would like to note that their applicability should be limited to the context of transfer capture alteration or use by non-state actors in which the state jurisdiction and compliance with IHL may be compromised. We would like to urge delegations to strengthen this framework in order to factor in those circumstances. My delegation feels that it is important to encourage states to develop legislation which would mean that laws can be developed and acquired in strict compliance with international humanitarian law. And we cannot content ourselves with an internationally legally binding instrument. It should also be complemented with objective national legislation that prevent impunity when it comes to the assignation of responsibilities. Finally, we would like to mention that we support the proposal made by Bangladesh in which human responsibility is promoted and not just responsibility over the machines with a view to enhancing accountability and responsibility for parties in an armed conflict. Thank you.


Chair [1:08:51]:

And Colombia was the last delegation on my list. So this is then. I think we have indeed exhausted the list of speakers with regard to box number five. Looking back at our discussion, we had a proposal indeed to move some of the text to different boxes. I have reacted to that earlier, but of course, we take note of the suggestions also to keep the elements in a logical order. There were made a few proposals also to merge certain powers or delete. Other delegations showed appreciation for that text. So there we will have to carefully look at how to remain the text in balance. Also with regard to box five, I think a number of delegations raised the issue of possible duplications in the text, um, and there were some, um, uh, suggestions with regard to streamlining and indeed, uh, a more, uh, uh, substantial one also with regard to, uh, clearly separating, uh, responsibility and accountability. Um, so this is, uh, the homework we take, uh, from this discussion. And then, looking at the clock, uh, we are, uh, just beyond 11:25, so that means we have still one hour and a half, and, uh, as I stated before, we have to make the best use of the time, um, available. we have available only 10 days a year. So I would like to suggest indeed that we now start addressing what I would call the homework we sent out yesterday at the end of the session with regard to further work, I would like to call it, on box number one. So I do hope you have all received this, uh, this text, and if not, I think it's on the UNODA webpage, as always, so accessible, uh, then also. Um, let me give a brief introduction, um, with regard to what we sent you yesterday. I think we can say that we have had three, three and a half days almost of intensive technical and legal discussions behind us, and I would like to start by thanking you for your interactive and constructive engagement. And after first reading of the revised text, it's clear that several points of divergence remain, some of which may be more difficult to solve than others. I am nonetheless encouraged by the amount of convergence as well as your willingness for flexibility and creativity that I've witnessed over the past days. I would like to call upon you to continue in this spirit, and I'm convinced that this will enable us to get closer to delivering on our mandate. By now, you will have seen the revision of box one. It's on the screen, both in track changes and in clear language, which I announced yesterday and which is, as I said before, on the CCWI as you webpage also. You know the webpage. Before opening the floor for initial reactions to this revised text, I would like to introduce the main changes and the rationale behind it. First, I took up the various suggestions by delegations to move and/or merge the disclaimers contained in the paras 1 and 2. noting that those are actually applicable to all paras in Box 1. I believe this to add accuracy while clarifying the content of the paragraphs. Second, throughout our discussions, we have heard numerous exchanges on the terms integrated combination of one or more weapons and technological components. Concerns were raised on whether this would imply that all components must be physically co-located or integrated into a single weapon system, thus limiting the scope of the characterisation. On the other hand, we also heard concerns that technological components must be understood as being related to the weapon system to avoid covering non-related technology, including technology for peaceful civilian uses, in our characterisation. I believe both points to be valid. Ultimately, I consider the main issue of this understanding to hinge on the word "integrated," which I have now replaced with "functional." With the new wording, I believe we are now more clearly saying that these various parts are linked, but clarify that they do not have to be in one place physically. At the same time, it clarifies that the technical components are linked to a weapon system and are not to be considered as a standalone item. This was suggested to be fixed through the terms like "there" or "auxiliary," and while we can continue to consider this, I believe this revised wording may already address that concern. But, of course, I would like to hear your views on that. Third, or whether to use the term user or operator. We have heard concerns that the term user may be too ambiguous and ultimately cover individuals that may be involved in the overall chain of command and control, but without the actual capability of intervening or otherwise affecting the functioning of the system. In response to this, we understand the suggestion to change the term to "operator" to anchor it more clearly in military and legal language. That said, the compromise here was to retain the term "human". In my view, this is key to ensure we are not talking about any other form of operator or operation that may or may not involve a human, and beyond this, and I dare say, Most important is that our work, after all, is on preserving the human element. As such, I do not think we should go forth with a characterization that does not mention human anywhere. With regard to those advocating for the usage of terms like a responsible chain of command and control, I would like to note that this concept is further elaborated later in the text, thus allowing us to keep this box focused. The same applies to concerns of responsibility and accountability, which are dealt with in box five, which we actually did just a few minutes ago. Fourth, only lethality. Here we heard several suggestions to remove and rephrase parts of the second para. In the end, I decided to remove the para two's first sentence. and to retain its second sentence as a standalone para with some changes. Considering our work is on lethal autonomous weapons systems, I do not think it's necessary, nor is it in our mandate, to find wording that defines how lethality includes death. It seems to me that this causality is very well understood. In fact, the initial addition of the explanation was not intended to explain that lethality means killing or causing death, but to ensure that other effects are still covered in our characterisation of loss. As such, I believe sentence 2 to be sufficient towards this end. In addition to this, examples of what effects beyond the loss of life may constitute were removed, not because I do not believe them to be crucial, but because I believe them to be more comprehensively subsumed in the revised wording. Ultimately, effects other than the loss of life can be understood to include both human injury or superfluous suffering, as well as the damage and destruction of objects. It will also spare us from entering into a debate in which everyone wants to see their specific examples reflected. Lastly, let me address what we did not change, and you perhaps expected me to change. The terms select and engage. I believe we are close to a common understanding regarding the critical functions. But so far, we have not been able to come up with text proposals to reflect that in the text. I think we will need some further in-depth discussion to find a solution for this. It is, therefore, my intention to have further discussions on this at a later moment. Having said that, I now welcome your views and comments on the revised Box 1, and the floor is open. I recognize the distinguished representative of South Africa. You have the floor.


South Africa [1:19:13]:

Thank you, Mr. Chair, for giving me the floor and for giving my delegation the floor. We thank you for sharing the revised box 1. While my delegation was unable to deliver or intervene earlier under this box, we have reviewed the current version and, building on our previous intervention during the last session, in this regard, we would like to offer the following comments. We believe that the revision represents a step in the right direction. And we understand the continued need for meticulous consideration on human involvement in the context of laws. In paragraph one, we welcome the progress made toward unpacking the critical process of target selection and engagement, which are central to the debate on meaningful human control. In this regard, we see value in the inclusion of the term "identify" in the text. We believe that this addition is important, as it reflects a broader spectrum of critical functions and contributes to a more inclusive and precise definition. Furthermore, the inclusion of the word "functional" in line 2 of this paragraph renders the definition ambiguous and open to interpretation. We therefore suggest that it be removed from the definition to enhance clarity and avoid misinterpretation. A clear and comprehensive definition not only enhances transparency, but also helps level the playing field. It empowers developing states to advocate for preventative norms rather than being confined to reactive defensive postures. Thank you, Mr. Chair.


Chair [1:21:04]:

I thank the distinguished representative of South Africa for sharing the views of her delegation. Next on my list is the Campaign to Stop Killer Robots, to be followed by Singapore and Mexico. But first, the distinguished representative of the Campaign to Stop Killer Robots, you have the floor.


Campaign to Stop Killer Robots [1:21:23]:

Thank you, Chair. Stop Killer Robots welcomes the update to the language in the revised box 1, paragraph 1, that characterizes an autonomous weapon system as a functional rather than an integrated combination of components. We believe this is a more clear and accurate way of formulating the concept being expressed there. With respect to the other refinements you have made, we do not see these as narrowing or restricting the scope of systems under discussion, nor do they alter any of the core aspects of the characterization of autonomous weapon systems contained in the previous text. Therefore, we remain of the opinion that the work on box 1 provides a sufficient basis to move to negotiations on a legally binding instrument to prohibit and regulate autonomous weapons systems that covers a suitably broad scope of systems and that could therefore adequately address the problems with autonomy in weapons systems that the international community has been discussing. Thank you, Chair.


Chair [1:22:23]:

I thank the distinguished representative, campaign Stop Killer Robots. Let me now pass the floor to the distinguished representative of Singapore to be followed by Mexico, Japan, and Switzerland, but first, Singapore. You have the floor.


Singapore [1:22:41]:

Distinguished Chair, my delegation would like to thank you and your team for your effort to revise the rolling text. We note the amendment to the chapeau and believe it appropriately draws attention to provide the characterization of laws that forms the basis for the elements of an instrument. It is with the significance of box 1 in mind that my delegation wishes to provide some observations and comments focusing on paragraph 2. While several delegations, including mine, understand "identify" and "select" to be separate and distinct functions, Other delegations consider target identification as part of or subsumed under the select function. Chair, in your opening remarks on Monday, you had indicated that you had chosen to include the term "select and engage" as essential characteristics of LOAs, with the assumption that the term "identification" is normally a precondition for these functions. Therefore, in the spirit of reaching consensus, my delegation would like to suggest a possible rephrasing of paragraph 2 as follows: "A lethal autonomous weapons system can be characterized as a functional combination of one or more weapons and technological components that can execute the functions of selection and engagement, which include identification, comma, without intervention by a human operator. And I repeat myself. The second part of that sentence, it would read, that can execute the functions of selection and engagement, comma, which include identification, comma, without intervention by a human operator. Of course, Chair, we will send the text to you in writing, and if we may, explain further. This formulation is a compromise proposal that seeks to better reflect the agreed precondition of identification as part of the function of select, while also addressing the concerns that some delegations have raised that inserting "identify, select, and engage" in a cumulative manner could be perceived as overly narrowing the definition. At the same time, this formulation also serves to meet the preferences of several delegations, including my delegation, to make clear that the identification function is a key task of LOS. With that, I thank you, Chair.


Chair [1:25:25]:

Let me thank the distinguished representative for Singapore also with coming up a new proposal with something we had not yet delivered text on, but this is certainly helpful for the discussion. Uh, I take, uh, a careful note of that. Um, next on my list, uh, and, and indeed, please send it also in writing, but I think everybody had the opportunity already to blot it down. Uh, next on my list is the distinguished representative of Mexico, to be followed by Japan, but first, the distinguished representative of Mexico, you have the floor.


Mexico [1:26:03]:

Thank you, Mr. Chairman. Once again, my delegation would like to voice its appreciation and thanks to you and your team, including the friends of the Chair, for producing this new version. The changes, I believe, are heading in the right direction because they are making the text more clear and direct. which is why I have a query, a question about the new wordings in paragraph 2 in section 1, box 1.

It reads in English, "The fact that laws can be used in a way that does

not result in loss of life." In this sense, that

causes confusion in my own mind. I don't know if we are changing the definition here. Should we use a different word from laws that does not result in loss of life? Because if we get rid of lethal, the word lethal, and elaborate on the definition, it appears that we are contradicting ourselves. We take out the lethal aspect, but we keep the word laws. Is that not a contradiction? Because laws by definition, I think, result in loss of life. So could I have some clarification? Because I at least am confused. I think we should refine that definition at any rate. Thank you.


Chair [1:27:58]:

I think the simple answer is here that loss is a loss also if there's not a loss of life. And we had a lot more words for that in the previous version, and as I explained at the beginning of this session, we tried to make that compact because we had already many different suggestions how to capture that. We had a long discussion at earlier sessions with regard to Indeed, the, the term lethal, but that's our mandate, that's how the group, uh, is named, uh, but, uh, of course, the, uh, the effects of the laws, uh, that is another issue, uh, and, and that has been, uh, addressed earlier in, in tax proposals, and this is, uh, this is a, a compact version of that. Um, next on my list is Japan, to be followed by Switzerland, Ukraine, um, and APILS. But first, the distinguished representative of Japan, you have the floor.


Japan [1:29:02]:

Thank you, Mr. Chair, and thank you for coming up with the revised version. On the chapeau, we find this amendment makes sense and welcome the change. With regards to the change to functional combination, we would be interested in hearing views of other delegations. As English is not our native language, we are struggling to find, understand whether it addresses concerns raised as well as whether or not it is creating new ones. On the issue of lethality, in our view, it is important to have a clear focus of our work in the GGE on lethal autonomous weapons systems, which is clearly set out in the mandate, while at the same time encompassing the scope of the work, which does not exclude injuries, superfluous suffering, nor damages and destruction of objects. We would prefer a more explicit reference to lethality, but we have heard views from delegations and understand the Chair's intention to accommodate various views in the room. Therefore, we are carefully considering the revised version. Next, on the issue of identification, we believe it is essential to have a common understanding on the essential characteristics of loss. In 2019, the GGE have agreed to mention identification, selection, and engagement of a target among the essential characteristics in its report. So a simple omission from the rolling text could imply that the group decided to disregard identification. This would probably not reflect our discussion. because as the distinguished delegate of Singapore have mentioned, we heard many delegations expressing its view that they consider identification as that is presupposed, is a precondition, or is part of selection function. Therefore, we support the proposal made by the distinguished delegate of Singapore to add, including identification, to be explicit of our common understanding. And lastly, we support the change of user to operator. Thank you, Mr. Chair.


Chair [1:31:31]:

Let me thank the distinguished representative of Japan for sharing the views of her delegation and commenting also on other proposals. Next on my list is Switzerland, to be followed by Ukraine, IPILS, Ireland, and Panama. That's what I have on my list now, and I will add India to that list. But let me first pass the floor to the distinguished representative of Switzerland. Sir, you have the floor.


Switzerland [1:32:01]:

Thank you, Mr. Chair. Our delegation considers that the work this week has clearly moved us in the right direction, and we would like to thank you also for guiding us in that direction. We value the constructive discussions and the careful changes made in the text, which have brought us closer to a workable characterization guiding our work. We also note that the changes, we value changes proposed in the chapeau, linking the characterization to these elements, and in paragraph 1. We note and valued the elements in paragraph 1 that have remained unchanged, such as the reference to "select and engage", which accurately reflect the thrust of our discussion this week. Allow me to say a few words on paragraph 2. Our delegation was among those cautioning against introducing additional language on lethality and be cautioned adding formulations such as with lethal force. The current formulation is definitely an improvement. We especially welcome the formulation included loss of life, which reflects existing IHL language. We note that earlier references to injury and to damage and damage to objects have now also been removed, references that we had very much welcomed. We understand your reasoning behind this step and we understand the new formulation as not excluding autonomous weapon systems that destroy or damage objects. In the spirit of cooperation and compromise, and with the awareness that the window for progress is closing, and for moving this text forward, clean up the characterization part, Switzerland stands ready to work on this basis and to engage constructively towards consensus. Thank you, Mr. Chair.


Chair [1:34:36]:

I thank the distinguished representative of Switzerland for sharing the views of his delegation and for showing flexibility with regard to tax proposals made. Next on my list is the distinguished representative from Ukraine, to be followed by IPAES, Ireland, and Panama, but first, Ukraine. Sir, you have the floor.


Ukraine [1:34:59]:

Thank you for giving my delegation the floor. The delegation of Ukraine is grateful to your Chair and your team for the outstanding work on improving the text, which we believe adds much clarity. Our only concern is that in paragraph one, the final words, "in execution of these tasks," quote, unquote, seem to be excessive and potentially confusing. It appears to us that this wording refers to the tasks of selection and engagement of targets. If so, we see no need in an additional wording after the comma. We therefore suggest deleting the words, quote, "in execution of these tasks," unquote, and putting a period in paragraph 1 after the word "operator." We have no further remarks and once again express our appreciation for the work being done to refine the text. Thank you.


Chair [1:35:54]:

I thank the distinguished representative of Ukraine for his comments with regard to in the The, the part at the end of, um, where is this, I know, yes, on the first power in the execution of these tasks, we, we had lots of discussions on that over the, the last year, um, one of the reasons we included there is to guarantee that it's not just a nominal human input, um, I think earlier I made the… I use the example of somebody has taken it out of the box and that was all he did. This is of course here, there is a clear link with selecting and engaging. So for that reason, we still have the words there. That's for my side. On the list, I have now the distinguished representative of APILS to be followed by Ireland and Panama. Sir, you have the floor.


APILS [1:36:59]:

Thank you, Chair. Many thanks to you and your entire team for the work that went into reviewing and balancing all the proposals and coming up with this revised text. We consider the inclusion of a chapeau that applies to all paragraphs of the box a positive development, which streamlines the text. We also have no concerns about the replacement of the word "user" with the word "operator". But we would like to make a few additional observations regarding the issue of a combination of one or more weapons and technological components. We were one of the delegations suggesting that the phrase "integrated combination" used previously was problematic because it might require the integration of all functions into a single device. Our suggestion was to insert the word functionally before the word integrated such that the text would refer to functionally integrated combination of weapons and other components. Completely replacing the word integrated with functional might, however, create new problems in interpretation as I believe the delegation of South Africa and Japan have already pointed out. The new phrase "functional combination of weapons and other components" could simply be taken to mean that the combination functions, that it works. We would reiterate our suggestion to use the phrase "functionally integrated combination". This would make it clear that the weapons and other components function together as a system. In other words, functionally integrated would mean that the weapons and other components are integrated in terms of their functionality, but not necessarily physically. Finally, we very much appreciate the revised and simplified language that you have proposed in paragraph 2. We are also grateful for your persuasive explanation for the removal of the first sentence of that paragraph. Thank you, Chair.


Chair [1:39:14]:

Let me thank the distinguished representative of APILS, uh, also for, uh, further elaborating, uh, previous suggestions and, and, uh, indeed, uh, giving us some background, uh, we will certainly give it careful consideration, uh, and I would also like to invite others to, uh, to, uh, reflect on that. But first, we have now the distinguished representative of Ireland to be followed by Panama and India. But first, the distinguished representative of Ireland. You have the floor, sir.


Ireland [1:39:47]:

Thank you, Chair, and to your team for the revised characterization section, which we are overall happy with. In paragraph one, we understand some of the ambiguity that the word "integrated" may have introduced, and we think that the word "functional" provides more clarity. We also would have been okay with just leaving it as a combination of, but we can live with either formulation. And we do think that the suggestion made just now from the previous speaker does have some merit and deserves some consideration. We welcome the retention of human in the last line. We're flexible between user or operator with a slight preference for operator, but for us, preserving the human element is essential in this paragraph. On the changes to the lethality paragraph, we think these changes add clarity and conciseness while still delivering the purpose of this paragraph, which is to ensure that systems that can cause injury or damage to objects are not excluded from our characterization, and so we thank you for this. On the identify issue, our preference is to maintain select and engage as it reads now. We thank our colleague from Singapore for their compromise proposal. Our initial reactions to the suggestion for now, it seems that the proposal still presents the issue of the cumulative approach, which is an issue for us and many other delegations. We are happy to look at their proposed text more closely. We think that there perhaps we can insert a may before the identify in their proposal to again to avoid the cumulative approach. and but we do also think that the original idea of including an and or could be a more elegant way to to um to solve this overall we think the characterization strikes a good balance that does not prematurely exclude a range of systems from its scope and closes potential loopholes and we just want to reiterate that these are indeed only elements of an instrument and further narrowing or finessing if necessary can and should be taken up during future negotiations thank you Mr. Chair.


Chair [1:41:59]:

I thank the distinguished representative of Ireland for commenting on the proposals made and on the proposals from the Chair, and also, indeed, for reminding us what we are doing here. Next on my list is the distinguished representative of Panama, to be followed by India and the Republic of Korea. But first, distinguished representative of Panama. Madam, you have the floor.


Panama [1:42:25]:

Thank you very much indeed, Chair. I'd like to add my voice to those who have thanked you for your work, you and your team, in an endeavour to find points of convergence on box 1. And following on the discussions we had on these paragraphs in the last few days, generally speaking, we agree with several of the changes made, such as, for example, in the chapeau. We can also be flexible with the changes in paragraph 1 pertaining to the replacement of "integrated" by "functional". We are flexible, either will do for us. And also, "functionally integrated" would also be acceptable to us, we are flexible. And as to human operator, we're flexible. As Eilidh just said, it would be important to state clearly that we're referring here to the human element. With regards to the critical functions, we would prefer to keep the paragraph as is, namely, select and engage, keep select and engage in the light especially of your analysis in your background paper in CRP.1, paragraphs 17, 19, and 20 in particular. And given that the inclusion of the word identify in the current wording could suggest or lead people to understand that it is cumulative, a cumulative requirement, you need to have all three fulfilled, we would therefore prefer select and engage full stop, so that the scope of the paragraph doesn't become too restricted. On 2, now, we welcome the attempt to increase the clarity thereof. However, we agree with what was said by the distinguished delegation of Mexico earlier on, namely that it might lead to confusion. We would prefer the previous paragraph, which is more of a compromise solution, wherein we explicitly referred to the lethal aspect, the lethality, but also it included other elements, meaning that was included not only death, but damage, destruction, objects damaged, and injuries caused. If I'm not mistaken, during the discussion, France had proposed the elimination of the first sentence and keeping the rest, perhaps that would be an option which we would be in favour of. But this being said, we would be happy to hear what other people think about these matters. Thank you.


Chair [1:45:39]:

I thank the distinguished representative of Panama, and I would now like to pass the floor to the distinguished representative of India. You have the floor.


India [1:45:49]:

Thank you, Chair. At the outset, allow me to appreciate your able stewardship of this group and your efforts, as well as those of your team and the secretariat, to accommodate different and divergent views and reflect them in the rolling text. We have taken due note of the edits made in box 1. We appreciate the edit made in the chapeau. and the reference to a human operator instead of a user. However, we are a bit disappointed that the text in its current version does not address the many concerns raised in the room regarding the characterization of a weapon system, including our own. We also note that the reference to lethal in para 2 has been removed, a formulation arrived at after lengthy discussions. Presently, we risk moving farther away from consensus. Since box 1 forms the basis of the discussions that follow in the remainder text, it is important that there is no room for ambiguity here. For the reasons mentioned in our previous intervention, we find merit in the edits proposed by Singapore, South Africa and Japan to include "identify" as a distinct critical function. On our part, we are flexible on characterizing laws as either an integrated or a functional combination as suggested in the revised version of box one, though we are of the view that it needs more discussions. Further, we propose the following edits. In para one, we suggest addition of the words, "they're integral" to qualify technological components to introduce clarity that technologies as they function within a weapon system are being referred to here, and technologies by themselves are not being covered. We also recommend explicit inclusion of design to identify. We have had a lengthy discussion on this before. We need clarity that only a system exercising independent judgment is being characterized here. By including design, we link the characterization to weapons that are designed to kill while preventing overreach into dual use technologies. Since the reference to lethal has been done away with in para 2, we also suggest addition of to cause loss of life at the end of para 1. Therefore, para 1 will read as under. A loss can be characterized as an integrated combination of one or more weapons and their integral technological components that are designed to identify, select and engage a target without intervention by a human operator in the execution of these tasks, to cause loss of life. In our view, the above edits will introduce the balance lacking in the revised version of box 1. We will also send these edits to you in writing. I thank you.


Chair [1:49:01]:

I thank the distinguished representative of India for her comments and suggestions made. Next on my list is the Republic of Korea, to be followed by Brazil and Sweden. But first, distinguished representative of the Republic of Korea, you have the floor.


Republic of Korea [1:49:18]:

Thank you, Chair. At the outset, I would like to express my delegation's appreciation to you and your team for hard work in providing revised box 1. I will make three points regarding this new text. First, overall, we are afraid that the revised box 1 does not address the core concerns raised by many delegations in this room. We do not share the view expressed by some delegations that this text provides a solid basis for further negotiation. without prejudging the formal negotiation that may take place in the future. Our aim is to build consensus around the set of elements. Regrettably, we are concerned that this revised text is not moving in that direction. In Korea, there is a saying that we sometimes take a one step backward in order to take step two steps forward. We hope that our exercise and discussion on this revised box 1 would take us to step forward, two steps forward. And I share the concerns raised by our Indian colleague that we are moving away from consensus. Second, however, my delegation is reassured that the Chair intend to continue discussions on the important and critical function of identification and try to find language that can address the concerns that we raised in regard to the omission of this function. In this regard, we thank Singapore for suggesting creative compromise language which we are open to considering. We reiterate our view that identification is a distinct and essential function separate from selection and should be reflected in the text. The Singaporean formulation can be useful in this regard. We again emphasize that our intent to include this function is not to unduly narrow the scope. but instead to provide the technically sound and practical characterization. I also find it difficult to fully understand the cumulative concerns raised as many delegations consider identification to be a sub-function of selection, then by definition those functions are combined. Third, we reserve our right to comment further on paragraph 2, the new paragraph 2, and the changes made to paragraph 1, but at this stage we wish to register our preference for the original paragraph two. And I will come back to further comment on the changes in paragraph one. Thank you.


Chair [1:51:43]:

I thank the distinguished representative of the Republic of Korea for stating the position of his delegation and making proposals. Next on my list is Brazil, to be followed by Sweden, Austria, and the United States. But I will now first pass the floor to the distinguished representative of Brazil. Sir, you have the floor.


Brazil [1:52:03]:

Thank you, Mr. Chair. My delegation appreciates the revised formulation that you have circulated. Overall, we believe that the proposed adjustments to the language are very positive and take us closer to an agreeable text. On paragraph one, we, beginning with the chapeau, actually, We believe that the reorganization of the text and the inclusion of a chapeau makes the language flow better and the reference to the context of the CCW apply to the whole subsection. So we welcome that change. On paragraph 1, we believe that the inclusion of the term "functional" is an important addition. It avoids the possible understanding that this combination of weapons and technological components needs to be physically contained in one single device. As we have discussed many times over this session and in the March session in particular, there are many autonomous weapon systems that do not necessarily fit this definition, but are autonomous in a functional sense. This includes close-in weapon systems that do not have detection capabilities, but rely on the detection capabilities of the platform in which they are installed. This includes swarms in which each component may have different functions but operate as a single whole, so this change is very welcome. We heard the proposal made by the representative of APILS, which would be to to include functionally integrated, that also works. Our delegation had proposed that formulation at the last session, but we are flexible as regards both. Regarding the replacement of the word "user" with "operator," we heard the arguments and the explanations that were offered over the course of this session and before, and we believe that this is possibly a step in the right direction. On paragraph 2, we take note of the changes introduced. While we welcome the simplification of the language, and in particular the fact that it steers clear of offering a definition of lethality, which would be a potentially perilous avenue to follow, especially given the possible impact on other instruments, we would have preferred to retain a specific reference to the damage to or destruction of objects or to cause injury, which we believe are essential elements of the systems that we are addressing here. We might be able to accept this language as long as there is a shared understanding that this does not exclude laws that were designed exclusively for anti-material purposes. And with that, Mr. Chair, I thank you.


Chair [1:55:14]:

I thank the distinguished representative of Brazil for stating the position of his delegation and commenting on the proposals. Next on my list is Sweden, to be followed by Austria, the US, and the ICRC. But first, the distinguished representative of Sweden, you have the floor.


Sweden [1:55:35]:

Thank you, Chair. From Sweden's side, we also want to thank you and your team for the efforts to find compromises and bring us forward. I can be very brief. Sweden welcomes and can support the new proposals in box one. And we continue to believe that we have a sufficient basis to keep us moving forward under your direction. Thank you.


Chair [1:56:00]:

Thank you for this very brief statement. Next on my list is Austria, to be followed by the United States. But first, distinguished representative of Austria, you have the floor.


Austria [1:56:13]:

I will take some of Sweden's time for our own intervention just to get a little bit deeper into the discussion and also, unfortunately, to make some general remarks again, because there was the notion that what we are achieving here mostly is consensus. And consensus is, of course, important, but there is one thing that is probably more important than consensus, and that is quality. So if you are adopting something just for the sake of achieving consensus and it's a low-quality definition in that case, I think we would not do anyone any favor. We would actually damage the discussions on this topic for decades or even centuries. I think this is something that we have to take care of with the utmost. We have to be very careful in how we move forward on this topic. If we go into a way forward where we have, and this is Probably for many of the delegations here in this room, including ours, the main concern is that we are adopting or working on the basis of a definition or characteristic that is too limiting, that excludes certain weapon systems. We understand the concerns. I mean, nobody wants to have an instrument that is too broad, that regulates things that are unnecessary to regulate, that don't fit definition exactly. This is also something that we have to balance out. The problem that we are seeing, if you look at specific systems that are out there, that many of them work similarly to non-autonomous system in the way how they are automated, but at the same time they have a high level of sophistication and also autonomous functions based on artificial intelligence, which actually is what creates the problems. But the theory, the idea, and the concept in some cases might be similar to existing systems that per se should not fall under this characterization. So as explained already earlier this week, for us, in doubt, the broader characterization is better, especially since those other systems would probably not be touched by those parts which come later in the text, which specifically talk about the systems, how to, which ones are prohibited and which ones are regulated. They will probably fall through the regulation part since they are too simple and not working in a way that is defined there. Just to make this general remark again, Another point I also have to make is that we are also against any kind of definition that is trying to limit the scope of our work altogether. We have more than 10 years of work here in this room, and if we limit the scope of our work in hindsight into something that is unnecessarily narrow, I think this is a real problem and we would not understand the motivations behind this, and we might have to question this motivation. If we are going into an extremely narrow definition that actually only talks about anti-personal weapon systems, or that only talks about science fiction, that probably will never come to life. And this is a problem. This is something that will damage our work, and is something that we cannot accept. Coming now to the paras, we can, as many other delegations, I think we can live with the changes in para one. I think functional is something that is actually an important and useful addition for some of the reasons that I have already explained, since it's a problem that some of the weapon systems can even move in and out of our definition and can be used sometimes as a autonomous weapon systems and sometimes not. So that we have a functional approach in our definition is actually to be welcomed. The select and engage and identify, I mean, we are open to a solution. I think we need one, otherwise we will not move forward. But something that is actually called the cumulative approach, which means that they have to, that there is like a checklist that a weapon system has to fulfill to be a weapon system, while at the same time being highly autonomous, I think that would not work for us. So even the name cumulative approach raises eyebrows and concerns in our delegation. We are also flexible with operator. I mean, user was something that we used for a long time, but operator can also be an alternative for that. Then coming to para 2. I think the biggest problem, I mean, the purpose of this para was actually to make clear that the lethality, which in our view is just a legacy of past discussions to actually get the discussions in a GTE off the ground. That's how it even came into the title of our working group. It is not the key aspect of our mandate as many claim. And that also was testified by the fact that in the past years, it was maybe a five or six delegations actually insisted on keeping it as a main characteristic of anything. This is now reversed now, and it's interesting that this is a phenomenon that we had to witness. The idea of this para and the idea behind the proposal by the ICRC was actually to clarify that the lethality does not limit the weapon system into something that is not purposed or designed against antimaterial. as Brazil has just explained. And this should also be the focus of this sentence, that it's very clear that the definition of laws as set out in para. 1 is not unduly focused on weapons which are designed to be used against humans, but all weapon systems. And this is an understanding, as we have explained many times, which is in line with the understanding of weapon systems in IHL in general. So if we are moving off that understanding and have an undue focus on the loss of life, which is an absolute novelty, and it should be an outcry of all the IHL lawyers in this room that we are going into that direction. So we would prefer that we actually have this explanation, that we have a broader understanding of weapons and not just anti-personal weapons, because nobody actually wants to ban just those weapons, and this has never been the focus of our discussion. So we prefer the previous formulation that is more clearer, such as to damage or destroy objects or to cause injury, as it is just the broader and more clear formulation, and thus avoid misunderstandings and actually a way forward where we are having different understandings what is the scope or what is actually our work. And this brings me back to my first point This is not the place to change any scope of our work, it's absolutely not this, and it's, it's problematic that, that this might be a motivation behind this. I thank you.


Chair [2:03:42]:

Let me thank the distinguished representative of Austria for commenting on the text, mate. I, well. It's your appreciation, of course, but I think I've been clear in what we have tried to achieve with this text and also with regard to the scope of our work. Uh, I think I said also in my opening remarks that, um, uh, we have three years of work, uh, and we want to do a, a thorough job, uh, so, um, I think that this is another honest attempt to do so. But it's, of course, free for delegations to have their own appreciation of that. Next on my list is the United States, to be followed by the ICSC. You have the floor, sir.


United States of America [2:04:35]:

Thank you very much, Chair, and good afternoon, colleagues. So, Chair, just to say, it's always hard to be the delegation following the delegation of Austria, but just to say, We actually believe that we need both. We need quality and we need consensus. And our delegation is confident that we can achieve both. Chair, we wanted to start by thanking you for the updated text. We appreciate our Korean colleagues' concerns in the sense that we may be moving backwards but sometimes you need to move backwards before you can move forwards. I think for those of us who have been at this for a number of years, that's definitely a feature of this process, but we do make progress, and I think the balance of our work tilts towards progress, and so we very much appreciate this latest proposal as part of that broader process. Chair, we We welcome the adjustments that you made to the chapeau. I think this largely responded to concerns that a number of delegations, including our own, raised about the need to orient the entire box around these two concepts. Chair, on the suggestion of replacing integrated with functional, here I think it's a starting point. We're not sure we understand what functional means. I know this was terminology introduced yesterday by the representative from APILS, but as the representative from APILS himself explained, this was actually not his proposal. And we were grateful for his further explanation as to how he had conceived the proposal. But the word functional is at least as a term that we've been using in this room, less than 24 hours old, we're not quite sure we understand it, and so we'd like to better understand it. We do think that the recommendation that was made earlier in this session by our colleague from India to include their integral just before technological components may be the elegant solution here. If we were to do that, you could delete functional. And it may be, hopefully it's the case that by shifting the idea of where the integration, you know, is occurring, that this will help those delegations who, like the delegation of Brazil, are concerned that The reference to integrated combination suggests that there may be a loophole. We actually don't see the loophole. Technologies that are integrated into a weapon system don't necessarily mean just those that are held within the confines of the physical space of the weapon, and so would encourage other delegations to think as we are now thinking about the delegation of India's suggestion to shift the reference to a bit later in the formulation. Chair, we welcome the suggestion, again, made by the United States, among other delegations, to shift from user to operator. We continue to have a concern, though, with the reference to human before operator, as we pointed out yesterday. No IHL instrument does this. By adding the term, we could imply that there is a category of non-human operators of laws, which we think opens the door to an ambiguity that we don't think we need to open. And so we would prefer not to suggest that there's some category that exists. So we would prefer to delete human and just limit it to the term operator. Chair, our most significant observation of related to the changes that you made is with respect to the end of the first paragraph. We would suggest, let me let me start with the edit that we would suggest, and then I'd like to start with the edit so that we can then explain what we're trying to accomplish with the edit. We would we would suggest at the end having the paragraph read, in the execution of these, and here we would insert, and other, so these and other tasks, and then we would add, necessary to initiate and execute an engagement. So let me explain why we're making this suggestion, because it's actually designed to address several different problems that we hear this morning and that we've heard throughout the last few days. So first, we continue to hear concerns raised by a number of delegations that the text is still not clear enough on the question of whether and how identification should be understood. As we indicated in our working paper, Our delegation believes that this characteristic can be understood to be covered by select, but not every delegation at this point seems to be comfortable with that. And so to the extent that others may not share this view, the edit is designed to more specifically allow for other tasks to be among the essential characteristics, and that could include identification. So we hope that that might help those delegations that have that concern. Second, Chair, the formulation that you've suggested to us today does, in our view, introduce a new level of ambiguity on the question of when the relevant intervention by the human should be considered in relation to the use of the laws. And you'll remember In previous sessions, we've had a lot of discussion about what the correct temporal frame should be. And we've previously raised the concern that if there's ambiguity on the question of timing, it may be possible to argue that an intervention in the design or the development stage related to these tasks could allow a state to argue that a weapon is not a LOS. even if it operates to identify, select, and engage the target after it's activated and without further intervention by the operator. And here I agree with the concern raised by the delegate of Austria, that if we spend all of this time defining characteristics in a way that allow states to argue that nothing is covered, That obviously would be an absurd result. So that's one of the concerns that we now have with this new framing. It's on the ambiguity as to timing. But this variation introduces another ambiguity that goes in a different direction. And that is that if the only two tasks that the machine performs without operator intervention or selection and engagement That means that this definition could cover any system where the operator performs other functions like identification or directing the machine to initiate the engagement. And so there, without the temporal framing, you could result in an overly broad formulation. So I think we've potentially achieved something that is almost unthinkable that there's a level of ambiguity that could allow states to interpret it both as covering nothing and covering everything. So from our perspective, it would be useful to try to tighten this up. And so when we had addressed these problems previously, we had thought that the simplest way to address this would be to add the phrase "after activation." And that, I think, continues in our mind to be an elegant solution involving just two words that would solve both of the problems, the overbreath problem and the underbreath problem. But recognizing that that was a suggestion that was controversial in the past, this morning's suggestion, which does approach it a little bit differently, is designed to get this temporal question in a way that would address both of these issues. Thank you, Chair.


Chair [2:14:32]:

Thank you very much for your extensive reaction to the text proposed, but also to the proposals made by other delegations. Just to have it correctly in writing with regards to the concrete proposals you made, Um, you would now like to propose the end of the first power, could I have the text up, please? Yeah, thank you, Mark. Well, we now have in the executions of these tasks to change into in the executions of these and other tasks necessary to execute the engagement, was that your proposal?


United States of America [2:15:16]:

Thank you, Chair. Almost one slight difference. Insert and other, so it would read execution of these and other tasks necessary to initiate and execute an engagement. So that gets at the beginning and the ending. That's the temporal frame issue that we're trying to focus on. Thank you.


Chair [2:15:52]:

And there you would initiate, you would in a way have what was initially, I think it was by France proposed after activation. You have already there, I think. Yeah, sorry to make this a bilateral thing, but I think we are all interested in better understanding and, uh, yeah, I need some time for that, but it is helpful as, as we have the best possible understanding now and for further discussions that might be helpful. So if you could, could react to that, it will be very helpful. If you need more time for it also, uh, we can get the floor anytime later, but, um, I think it's useful for us all to know that.


United States of America [2:16:48]:

Thanks, Chair. So we did multilateralize this a little bit. We've consulted with our neighbors. So I think one, so again, after activation, as suggested by France and the United States in previous rounds, we think would solve the problem. But note that after activation was related to activation of the system. Here, we're focusing on a conceptually slightly different way to think about it, was the initiation and the execution of the engagement by the system. So I know there's some nuance there, and we're happy to further discuss it on the margins later. But again, either one of these, I think, would be solutions. But we are trying to offer maybe a new way of thinking about it that since after activation did not seem to enjoy consensus, perhaps this one would.


Chair [2:17:48]:

No, thank you for that. But indeed, this, we consider that already to be also the temporal element then. That's very helpful. We still have some time. And Next and not least on the list is now the distinguished representative of the ICRC. You have the floor.


ICRC [2:18:12]:

Thank you, Chair. On paragraph 1, we see the current terms of select and engage as having a more common understanding. Echoing the comments just made by the United States, we also see identification as being covered by the select function. However, we have listened to the proposal by the distinguished representative of Singapore with interest. and we think it reaffirms that we are all aiming to cover the same functions, but we are using different terminology. Our worry is that the inclusion of the term "identify" could be interpreted to cover the mere pre-setting of targets, a process often done by humans prior to activation. To emphasize, a weapon system should fall within the scope of an instrument, even if the identification of a type of target is carried out by humans prior to activation. The key point is to capture all systems in which a human does not select the specific target and does not engage that target with force. However, if high contracting parties consider it more appropriate to refer to both functions separately, its retention could be considered in the spirit of working towards consensus. provided that a qualifier is included to clarify that the notion of identification refers to potentially finding and matching potential targets against the generalized targeting profile to determine whether they should be engaged or not, and not the mere pre-setting of targets or pre-selecting of targets. We also support Belgium's proposal to bring back the and/or formulation. So it would read, identify and/or select and engage. Moving to paragraph two, we echo the comments made by the distinguished representative of Austria that this paragraph was indeed originally proposed as a compromise to keep a reference to lethal without narrowing the scope of the characterization. This was particularly the case since it became clear throughout the previous GGE sessions that delegations do not all understand the term lethality in the same manner, which would make its inclusion without a qualifier problematic for reaching a common baseline. That said, we are fine with the deletion of the first sentence, also for the reasons you proposed, Chair. We think this could also resolve the issue of defining lethality in a non-lethal manner, and address the tautological concern raised by some delegations. However, on the second sentence of paragraph two, we are concerned that the current phrasing could be read as requiring a system to have the capacity to cause loss of life, which could exclude many anti-material autonomous weapon systems or weapons which may not be capable of causing death, but are capable of causing unnecessary suffering and superfluous injury, or of damaging or destroying objects. We don't think this is the purpose or intention of delegations today, particularly given that many of the prohibitions and restrictions that appear later in the text do cover objects. As a possible compromise, we could suggest wording along the lines of the fact that the laws can be used exclusively against an object, or does not result in loss of life, does not exclude it from this characterization. I thank you, Chair.


Chair [2:22:25]:

I thank the distinguished representative of the ICSC for commenting on the text proposals and for proposing some new text, which I read as the fact that can be used exclusively against an object or does not result in a loss of life, that will be the change.


ICRC [2:22:50]:

Yes.


Chair [2:22:51]:

Thank you very much for that. Yes, we still have some time. Next on my list is Portugal, to be followed by Spain, Australia, Turkey, but I don't think we come that far. We then have the Russian Federation, Belarus, Canada, Israel. But let me first pass the floor to the distinguished representative of Portugal. You have the floor.


Portugal [2:23:15]:

Thank you, Chair, and thank you also for this revised version of box one, which is very welcomed by my delegation. It is a very honest and, I think, accomplished try to bring on board different views, which is not easy. So we would like, first of all, to congratulate you for that and to provide our overall support for your text. Having said that, we have some minor comments. First of all, we would like to support inclusion of the term "functional," although we are also eager to consider other options. We heard "functional integration," which could be something also -- or "integrated." That's something that we could also go along with. We also welcome the change of the term "user" for "operator." Although, as was also mentioned by others, including Ireland and Panama, for us, the main element here is really the reference to human. Then, regarding the introduction or not of the term "identification," we are also of the view, as others, that it is already encompassed in the term "selection." But again, we are open also to consider other options. Finally, in what concerns paragraph 2, we have some doubts and concerns about the removal of the reference to lethal and lethality. We do understand that this is a compromise, of course, but we quite frankly, we would like to, we would prefer to retain the reference to lethal, because it's indeed, like it or not, it's a term that is in our mandate. It is a term that characterizes specifically the type of weapon systems that we are considering, at least it is under our purview, but also because there are other discussions outside of the GGE laws that also take differently, but the approach is different when if the lethal element is there or not. So I think also outside for purpose of the discussions outside of this group, it would be important to retain lethal. Also in terms of the legal discourse, it would be also make sense to define lethality or lethal and then clarify that it includes specific, what does it include specifically and what doesn't. Of course, we are, and having said that, we are minded that this is not a negotiation of an international instrument, hopefully, That will be the next step, but we are not yet there. And therefore, and although we have our reservations for reasons that were also mentioned by others, including most recently the ICRC, we would prefer to retain a draft closer to what it was in paragraph 2. Nevertheless, and also taking in mind efforts that are required to reach a consensus, If that is the majority of the views here in the room, we are happy to go along with the text as is. Thank you again.


Chair [2:27:00]:

I thank the distinguished representative of Portugal for commenting on the text proposed. Next on my list is Spain, to be followed by Australia. But let me first pass the floor to the Ambassador of Spain. You have the floor.


Spain · Ambassador [2:27:16]:

Thank you very much indeed, Mr. Chairman. May I begin by thanking you for this new text and thank also those delegations who have, I believe, made very constructive proposals, which are real food for thought. What I'm going to say cannot respond to all of the proposals, which, as I say, have been very interesting, but we can circle back to them subsequently. We support the changes, Mr Chairman, made in the chapeau of this revised version. In our view, they are fitting, very fitting, and on paragraph 1, three or four comments echoing those made by other delegations. With regard to the words "functional" or "integrated", we like the word "functional", which you have proposed. Many delegations have explained the importance for it, because some instruments are integrated and some are not. Therefore, "functional" works better, because it enlarges the concept somewhat. The concept of functionally integrated or similar concepts similar to functionally integrated are interesting to us because the components that are physically integrated to the systems as well as those that are not physically integrated can be represented by this characterisation. That's why we feel it is interesting. Moving on to another topic now, it's important for us to keep the word "human" in this paragraph. Very important that we keep the word "human". And with regards to the change from "user" to "operator" at the end of paragraph 1, we've listened to, very carefully in fact, to your explanation for that change. The aim being to reduce the ambiguity which the use of the word user could lead to. On this change, let me approach this from a legal point of view. One of the purposes of international law and international humanitarian law is to ensure that states shoulder all their responsibilities stemming from any breach of international law, especially with regards to those persons who, through their acts of omission or commission, thinking of the user operator here, are guilty of that breach of international law, which means that we need to use a term which is broad enough to include the chain of command involved or not in this liability for this act. And for that reason, and because we didn't want to limit responsibility solely to the operator of the system, consistent with Article 86.2 of Additional Protocol 1, we felt it was more fitting to use the word "user". That is the legal explanation, the legal approach. At any rate, responding to the explanation and also responding to this need to show a spirit of compromise, despite what I have just said, from a legal point of view, we can support the change from user to operator, but we must keep the word human though. With regards to identify, select and engage now, we were happy with referring solely to select and engage. This being said, we found that Singapore's proposal was very interesting. made this morning, which has its merit is to make explicit the understanding that we assumed everyone had that select included also identify. And Singapore makes explicit that understanding. Also, we are also very much in agreement with Japan about the fact that if you get rid of identify, which contained a characterisation which we had made and which we had agreed by consensus in 2019, could suggest that we were not including this commitment anymore, which means that we support Singapore's proposal. but we are also looking very carefully and with interest at the US's proposal. So there is this attempt to return to identify, and around that attempt, we should identify or seek to achieve an agreement. Finally, in paragraph two, Spain would be grateful for us to refer more directly to and comprehensively to a lethality echoing what has just been said by Portugal just now perhaps by reverting to the previous wording this being said at the same time we think that we should carefully consider the proposal made by the ICRC this morning which brings to the text once again. I won't read the whole thing, because you all have it before you, but that proposal enlarges that concept in a way which we find very interesting. Once again, thank you very much, Mr Chairman.


Chair [2:34:15]:

Let me thank the Ambassador of Spain for commenting on the text proposed by the Chair and text proposals made earlier this morning. Next on my list is Australia. Let me pass the floor to the distinguished representative of Australia. Madam, you have the floor.


Australia [2:34:35]:

Thank you. Thank you, Chair. I will be brief, hopefully. So just to start off by saying Australia supports the creation of the chapeau. If we could have the text up, it does make it much easier to respond to the changes, just because there have been so many changes overnight.

Thank you.

So Australia supports the creation of the chapeau. We think this will serve well for the structure of this box. We have listened with interest to interventions today on the use of functional combination, integrated combination and functionally integrated combination. We're still considering the new term, the use of functional combination. Our initial view is that the proposal that we heard from APLS, that it should be functionally integrated combination is more correct technically and makes more sense, but we are very happy to discuss this with other delegations. Likewise, we are comfortable with the use of either the term "user" or "operator" in conjunction with the word "human".

The GGE has agreed that accountability attaches to humans, and from our perspective, this is reflected in both expressions.

Chair, we think there is merit in some of the proposals to be more explicit about the fact that autonomy occurs from the moment of activation until the completion of the law's mission, Laws operate autonomously after human activation to complete the critical functions. We would therefore support the suggestions that have been made to insert "after activation" behind "technological components" in para 1. Turning now to the definition of "lethality" proposed at paragraph 2, we can accept the revised wording. This simplification, from our perspective, retains the notion of lethality while recognizing that laws could have non-lethal effects. This will capture the systems of concern which sparked the establishment of this GGE.

Thank you.


Chair [2:36:41]:

Let me thank the distinguished representative of Australia, and she has been brief indeed, so I have time for at least one more. delegation, and that's the delegation of Turkey. Let me pass the floor to the distinguished representative of Turkey. You have the floor.


Türkiye [2:36:58]:

Thank you, Chair. With your indulgence, as I was shuffling between various disarmament meetings, I'll touch upon briefly on box four and five before moving to the new box one. As was pointed out by other delegations yesterday, I'm not talking about box four, there are states in this room who are not parties to the first additional protocol to the Geneva Conventions. Therefore, we agree with the assertion that it is necessary to adequately reflect varying legal obligations of the states in box four. We attach importance to maintaining the reference to, here I quote, "with their obligations under international law," unquote, in the first paragraph, and we're open to moving this to the beginning of that sentence. We also support the proposal to add this phrase to paragraph four. Regarding box five, we have similar observations for this box and attach importance to maintaining the applicability aspect of international law and IHL. Therefore, we strongly urge you to keep the references to applicable in paragraphs 2, 4 and 5. Same goes for the phrase in accordance with their obligations under IHL in paragraph 3. As we emphasized the other day, maintaining the word should in boxes 3, 4, 5 is important to refrain from prejudging the outcome of this group's work. Mr. Chair, thank you for sharing the revised formulation for box 1 last night. While we're still evaluating the second paragraph after the chapeau, allow me to already lend our support to the compromise proposal shared by Singapore this morning regarding the first paragraph after the chapeau. This is an eloquently crafted bridging proposal that deserves consideration, and my delegation will also take a look at the U.S. proposal. Thank you.


Chair [2:38:46]:

I thank the distinguished representative of Turkey. And now I'm looking at my watch. It's less than five minutes. I think I have the Russian Federation as the next on my list. Would you like to take the floor? Yes, we would like to hear your views very much. So you have the floor.


Russian Federation [2:39:06]:

Thank you very much, Chairman. We will try to say everything in those five minutes. First of all, we'd like to echo the words of gratitude addressed to you for the preparation of this possible compromise solution and for the efforts that you've made, along with your team, to find a compromise. Our delegation greatly appreciates this, and we are ready to continue to provide every support to you in this respect. Unlike some delegations, we believe that this is a step in the right direction, which is trying to strike a balance between different interests which have been voiced by delegations. That said, our delegation does also have a number of comments and proposals to make. First of all, regarding the chapeau. We would believe it would be correct to use for the second part of the chapeau wording from the overall chapeau of your document, where there's a reference not to the elements of an instrument, but to the group of government experts itself. And we agree. that such a definition or understanding regarding the characteristics will be made for the purpose of these elements, this group of governmental experts. To say that this is for the purpose of these elements of an instrument is a bit premature. We are not yet working on an instrument, and we are not talking about negotiations concerning an instrument. We are engaged in the execution of the mandate of the work group, which is quite broad in its formulation. To mention the purpose of these elements would be premature. Regarding paragraph 1, we are ready to consider the compromise wording "functional" instead of "integrated combination", "functionally integrated combination". We believe this would be a compromise that should be given consideration and which overall would reflect the essence of what we're trying to say, which is that we're not talking about the entirety of the system with all its components. The functionally integrated combination, functionally component. the functionally joined components in the context of a single system. And that is what we feel we are trying to talk about. As regards the technological components, We consider it not to be advisable to make any additions to this wording, with the exclusion of the word "auxiliary", which we feel would be appropriate, especially bearing in mind that we're talking about lethal autonomous weapon systems, and in this connection, the strike components are going to be of prime significance. There will also be other technological components that will be auxiliary in nature. That is why we feel that the word auxiliary is necessary before technological components. As regards identification. Our delegation has already said, and we are ready to repeat, that we are flexible on that. We are satisfied with the combination of the two functions, selection and engagement. We are happy to talk about three functions: identification, selection and engagement. However, it would be problematic for us if we were to return to the wording "identification or selection", because that would imply a theoretical situation in which a target could be identified without being selected. We would like to exclude such situations, as we consider that the selection of a target to be one of the key characteristics of a lethal autonomous weapon system. As regards the human operator or user, we're happy to be flexible. We're happy with both operator or user, human operator, human user. As regards lethality, we consider that it is necessary to include that characteristic in paragraph 1, as we are talking about lethal autonomous weapon systems. We are not in a position to change the mandate of the GGE, which covers lethal autonomous weapon systems. So we consider it would be possible to agree to the proposal made by India, which is and cause loss of life. Regarding the time factor, we insist on the need to delete the words in the execution of this tasks, bearing in mind that the characteristic functions for these weapon systems will be inherent throughout the entire life cycle of these weapon systems. And we do not consider it necessary to create a grey zone, as it will cover all of the life cycle. A grey zone regarding all of the other stages with the exclusion of the use. We also do not feel that it is necessary to come back to the words after activation. We felt that we'd moved on from that. And we would have liked to once again underscore that for us, it would be problematic to include that wording as it creates additional issues to which we feel we might not get an answer within this group. The second paragraph satisfies us as it is worded. The third paragraph, again, we believe it is necessary to more clearly formulate a consensus here so that it is clear that at this stage already, paragraph one and the descriptions it contains does not exclude a specific type of, does not include certain types of systems.


Chair [2:47:24]:

Let me thank the distinguished representative of the Russian Federation for sharing the views of his delegation and also proposing new text. Let me also thank the interpreters for their indulgence. It's very helpful that we were able also to hear his views. We will continue. I would like to propose that we will start in the afternoon at 3:30, so I need a little bit more time. But I think we, so far, because we still have, um, quite a list to go, um, interesting, uh, the reactions we had with regard to changing into operator, and I think the, um, the text I proposed as improved by, by, um, APILS, um, seems to be, um, uh, getting some, quite some support across the room, so we, we ha- clearly have positive, uh, feedback on a number of issues. Clearly, there's also still a lot to be done on other issues, and we might have also one or two new issues that would fit into what was said by, I think it's the procession of Aechternach, that you go one step back and then two steps forward again. I see the Luxembourg colleague already nodding. With that in mind, I wish you all a good lunch and see you back in the room then at

3:30.