I I I I Excellencies, distinguished delegates, could I ask you to take your seats again? Well, good morning, Excellencies, distinguished delegates.
Let me start with expressing my appreciation for the informal discussions we had yesterday evening.
We made some progress with regard to the characterization, but that was only some progress and we really have to step up if we want to deliver on the mandate to deliver a set of elements of an instrument and measures, et cetera by Friday.
I already announced yesterday at the end of the meeting that we will have another round of informal consultations tonight and that will be so indeed and that will be from 730 until 10:00 in the evening.
Now, with regard to today, the morning and the afternoon, I intend to walk with you through the rest of the text.
The text that was not yet discussed, we discussed para 25 and we discussed what was it 3530 8406 4083 5408.
Please leave that out.
By now, I think the chair but also colleagues in the room know your national positions on that.
So we would now like to hear the positions the other paras.
The tonight we will have an informal discussion for high contracting parties on the issue of human involvement.
O P 36, during the day, my team will still be working on the text to be discussed tonight.
Of course, you need to be prepared for that, so we will try to circulate it somewhere in the afternoon to give you some time.
For that reason, we also start at 7:30 to allow you to have some time to read it.
Also because I have some informals to do myself.
I intend to start the afternoon session not before 4:00.
So we will have a meeting.
We will meet this morning until 1:00.
We will resume then at 4:00 until six, and then we will reconvene in an informal setting at 7:30.
So that is with regards to the schedule.
I hope it's agreeable to all of you.
Now, I would like to invite all of you, as I said, to start considering the remaining powers of my additional suggestions.
You have seen that I've only additional suggestions as of 33, but you are, of course, free to share your views also on the powers 26-23.
With regard to 33, while the core meaning of this power is not contested, there continue to be divergent views on what constitutes the correct language for accurately reflecting existing obligations under IHL, as well as on whether additional precautions should be included to account for a potential change in the status of protected individuals and objects.
The suggested formulation retains the idea of the original wording while seeking to simplify and clarify the power and preserving existing IHL obligations and the understanding thereof.
Perhaps you can put the text of 33 and I also would like to recognize the concern that the suggested formulation could be understood as implying a lower level of protections for the civilian population.
However, I would like to reiterate that under existing IHL, both individual civilians and civilian objects may under certain circumstances, lose the protection afforded to them under IHL, while the civilian population as a whole cannot under any circumstances.
Let me now turn to Power 39.
At the request of several states, this power has been brought in line with the exact wording of additional Protocol one, Article 36.
The original formulation was intentionally kept shorter as the Chateau already refers to the life cycle of laws, thereby making explicit references to their study, development, acquisition, or adoption, which are all stages of that life cycle, somewhat duplicative.
Nevertheless, we must, of course, remain mindful of the importance of preserving agreed IHL terminology and should avoid any unintended changes to established language.
Let's now move to paragraph 40.
And it Spars, they were the focus of extensive discussions during our previous session.
In line with that, adjustments have been made to address some of the points that have come up repeatedly.
Spar Alpha, while there appears to be broad agreement that testing and evaluation are important to ensure an adequate understanding of the capabilities and limitations of laws, as well as the anticipated effects in different circumstances, questions remain regarding the scope and methodology.
Over time, the language has been adjusted to ensure that testing will be limited to simulations rather than instances where battleground will become a testing zone.
However, further change seem to be necessary to avoid unconstructive ambiguity, including what constitutes a realistic simulation.
Now on Spower Bravo.
This includes a stylistic change by introducing in before control to provide further clarity to the power.
While chain of command and control is, of course, well understood and agreed terminology within and outside the GGE, the addition of in is intended to refer more specifically to personnel who exercise control over the systems, who may not necessarily be the same personnel as those within the broader chain of command.
This formulation is now also more closely aligned with the language contained in P 44.
Now about subpara C.
Changes have been made to address concerns over ambiguous terminology, notably by removing adequate, as well as to remedy prior inconsistencies, including the omission of the need for not just the operation, but also the effects of the laws to be predictable.
This power, like others of the text certain other parts of the text, is one of the elements that has faced extensive exchanges throughout the various iterations.
I'm well aware that the terminologies used in this power remain problematic for some delegations.
However, let me stress that for others, these are an essential element and a crucial condition for the lawful use of laws.
Furthermore, this power must be read in conjunction with its chapobot.
I therefore ask delegations, when addressing this power to engage on the concepts contained therein, with a view to finding compromise solutions rather than remaining stuck in a discussion over whether it should or should not be deleted.
Now to paragraph 47.
This power includes now a small technical change to state more clearly that internal mechanisms relate to a state's own use of laws and not to address incidents of other states use of law.
It's laws.
Lastly, paragraph 48, this para has been moved down from its previous placement in order to avoid varying interpretations or confusion over which elements and measures it covers.
Sitting at the end, it now more clearly addresses the text as a whole.
In addition, it also has been brought in line with the mandate of the group.
Having said that, let me also draw your attention to a request from the interpreters.
If you have any statements prepared in writing, that you email them to speeches at un.org.
We are having a very, very complicated discussion, a very technical discussion.
I understand this is really a challenge for our interpreters also.
They love to do it, but it's not an easy job, and we need, of course, the best possible interpretation we can get in order not to misinterpret what some of us are saying.
We need to get it right because otherwise we have confusion in the room and that's not helping if we want to build consensus.
Having said that, let me now open the floor for interventions from delegations.
The floor is open.
I recognize the distinguished representative of the Russian Federation.
Thank you for breaking the ice.
You have the floor because I was on the point of proposing then that the chair will assume all the suggestions in the next version of the draft report.
But let's hear views of delegations.
You have the four.
Thank you, Chairman.
We think that this hesitation from the delegations, including from ours is caused by the fact that the quantity of paragraphs under consideration is quite significant and requires some additional consultations within the delegations.
Regarding the paragraphs you've proposed, we would like to make a number of comments.
On paragraph 33.
We don't have any comment regarding your suggested amendment.
We consider that it is a step towards a compromise and this wording could be accepted for adoption by consensus within the group.
Having said that, we still have concerns which we already had about the stylistic aspect of paragraph 33, specifically the use of the wording, it is prohibited.
We believe that this wording is not appropriate in the context of this paragraph, and we would suggest replacing it is prohibited to make must not be before the word the object of an attack by laws at the end of the paragraph.
Regarding paragraph 39, we'd like to express our gratitude to you, Chairman, for having taken into account the concerns that delegations voice, including our delegation, regarding subparagraph A of 39, and we are prepared to accept the proposal that you suggested in subparagraph A as worded by you.
But we do still have concerns about subparagraph B.
We feel that this little paragraph is superfluous because it is covered by the understanding of a legal review itself, which foresees following the procedure that's described under the circumstances that are mentioned in subparagraph B right now.
So we would prefer removing subparagraph B in full.
On paragraph 40, little A in its new wording, requires, in our view, some additional textural refining so as to make it slightly clearer and concise.
As a possible proposal to tweak the wording, we would like to recommend considering the following amendment, specifically adding after conducting testing and evaluation, the words to assess their performance.
Then after this sentence, we could word the ending of the paragraph as follows.
Including within simulations in operational environment.
Thus, subparagraph A would read as follows in full.
Conducting testing and evaluation to assess their performance, including within simulations in the operational environment.
Such In this case, this subparagraph would be more clear and precise.
And it would unambiguously be clear from the paragraph that we're speaking of conducting testing in the context of simulations within the operational environment.
We believe this is important to make clear.
In order to enable the best possible preparation of the experts who will be working with these weapon systems and also to ensure reliability of the systems themselves within their use.
On subparagraph C, our previous concerns remain the quite serious concerns about this subparagraph.
Specifically regarding the concepts that are contained in the paragraph, predictability and traceability, explainability.
We feel that these concepts are not clear in terms of how they're understood.
We still don't understand what exactly they mean and in particular, how they will be applied in practice.
As we have said previously, we would be ready to work on defining what is meant behind these concepts.
We could do that separately or in the context of this subparagraph itself.
But for us, one thing is clear without clarification as to the contents of the concepts that are contained in this text, agreeing to just mentioning them in subparagraph C would be problematic for our delegation.
On paragraph 47, of course, we are not against the amendment you've suggested.
However, there is a point that does raise a question.
We consider this paragraph to be an attempt to interfere in the internal affairs of states parties, given that we're speaking about mechanisms at the national level.
And the language that is in this paragraph is through its style, quite strict.
The word must is used.
And that, in essence, is like a dictate regarding the actions that state parties must take at the national level.
We would like to avoid any such wording in this document and leave the issues of domestic matters to the states parties themselves.
All the more as we have Paragraph 45, and paragraph 46, which foresee in general terms, the possibility of relevant measures at the national level that could be taken by states parties and that are also referred to in paragraph 47.
In this connection, we believe we must remove paragraph 47 in full.
On paragraph 48, which is a new suggestion, it's a rewording of former paragraph 36.
Our delegation continues to think that this paragraph has superfluous elements as we have paragraph 39, which specifically is about the stages of development.
And therefore, we feel that paragraph 48 is not necessary.
It is basically an overlap with paragraph 39.
So therefore, we believe it is necessary to remove paragraph 48 in full.
Thank you very much.
Let me fat to distinguish representative of the Russian Federation for sharing the Commons Office delegation.
Yes, with regard to the language it is prohibited, you now propose must not be the object of an attack.
Yeah, we have been playing with language there to not create new obligations, et cetera, but we have, of course, um We have here standing a ch, and for you, prohibit is an issue.
I think for others, it might be must again.
Verbs are something of a headache, I would say.
But we will look into that.
Also with regard to what you say about, um, 47 in relation to 45 46, we might get some inspiration from 45 46 to make perhaps the texts in 47 more palatable for you.
Next on my list is the distinguished representative of Skiller robots.
You have to fly.
Thank you, Chair.
Regarding the remaining paragraphs in the report, the subparagraphs of the paragraphs numbered 39 and 44 in the draft report address some of the core components of meaningful human control, which are that users must have an adequate functional understanding of systems and must sufficiently limit their operation and effects.
We welcome that the key concept of predictability, reliability, traceability, and explainability have been reinstated to the set of elements in paragraph 39 and 40 on your screen and that shares suggestions for further clarifications on what should be required.
These concepts have been central to the discussions on the means by which meaningful human control can be achieved and must be retained.
Points A to C of paragraph 44 and 45 on your screen address the need for users to have an adequate functional understanding of systems to exercise meaningful human control and ensure responsibility and accountability are upheld.
All these elements must be retained in the text to give a basis for negotiating international rules on maintaining meaningful human control and the use of force that goes beyond simply restating the text of existing law.
Regarding paragraphs 40 and 41 in the draft report now 41 and 42, having discussed and developed their understanding of the different types of bias that raise concerns when it comes to autonomous weapons systems, states should be sure to retain these elements as a starting point for a deeper consideration of the rules that are needed to respond to them.
When it comes to algorithmic bias and the reproduction and reinforcement of structural discrimination, such as sexism, racism, and ism, part of the international response must be a prohibition on autonomous weapons systems that target people.
Finally, we welcome the suggestion from the chair to move what was paragraph 37 in the draft feedb to the end of the text as paragraph 48 to clarify that the whole set of elements should be taken into account across design, development, and use.
Thank you, Chair.
I thank the distinguished representative of Skill Robots for sharing the views of a delegation.
And let me now invite other delegations also to share their views on what is on the table.
The text was circulated on the 5th of June, and my suggestions are from the beginning of August.
So it's not that you received it overnight.
Thank you.
Let me now pass the floor to a distinguished representative of Pakistan.
You have the floor.
Thank you, Chair.
Chair, on this section, we do recognize that there are a number of paragraphs, but let us first address the ones that in our delegations point of view are very important.
We think that paragraphs 39 to 48, they are important in the sense that tell us how a state makes itself capable of exercising what would be required by human judgment and control.
Therefore, this is how we read this section.
On paragraph 39, my delegation supports the Article 36 review obligation as it is currently drafted.
Uh, we would note and welcome also subparagraph B of 39, which extends review to further development and modification affecting the identification, selection, and engagement functions or the anticipated effects.
In our view, we think this is important and this matters more for the autonomous systems than for any weapon that has come before in history.
We also welcome the reference to voluntary exchange of practices with safeguards for national security, and we think the language is properly balanced right now, and we would not wish to see it disturbed in either direction.
Here, I turn to paragraph 40.
Paragraph 40 is for our delegation, the most important in this section and we support the retention of predictability, reliability, traceability, and explainability in this section.
In our view, as we have stated a number of times before, these are the conditions under which a human judgment about anticipated effects can be made at all.
Let me put it plainly.
Paragraph 36 asks human beings to determine on the basis of anticipated effects that use of a system will comply with IHL.
That determination is only possible if the system's behavior can be anticipated, which is predictability.
It is only meaningful if the system behaves that way consistently and which is reliability here.
Paragraph 44 requires individual accountability and paragraph 47 requires effective investigation.
These concepts of individual accountability and effective investigation are not possible.
They cannot be established after the system has operated if there is no traceability and explainability in built in here.
Therefore, all these four concepts serve distinct purposes in this section and we think that they are the load bearing connection between the requirements of control and human judgment and the accountability provisions which appear later on in the text.
Chair, we are also happy with paragraph 43.
We support this and we attach particular importance to the two elements that are there, and we would like to see them retained in the final text.
First is that the responsibility cannot be transferred to machines on which we have converged for a number of years and second is that final sentence extending this across the entire life cycle, that it's not the states which are responsible for their decisions regarding laws, but also for the consequences which follow from them.
That capturing that aspect is important also for our delegation.
Share briefly and lastly on paragraph seven, we think that this is a very balanced formulation.
It requires states to ensure effective internal mechanisms, and I would stress on the word internal being the operative word.
It does not prescribe any forum or procedure or institution or specifies detailed processes.
It leaves entirely to each state how such mechanisms are to be constituted.
So therefore, we do not understand any concerns on this language on the account being too intrusive in the internal affairs of a state on this.
And the duty to investigate credible allegation of serious violations of IHL is already well established concept in the customary international law as well.
These are our preliminary observations on this paragraph at this stage, but we can always come back again in terms of responding to any new proposals.
Thank you.
I thank the distinguished representative of Pakistan for sharing the views of his delegation and also some of the backgrounds of concepts we have been discussing for quite some time.
I think it's most helpful.
Next on my list is the United States to be followed by Turkey, but let me first pass the floor to the distinguished representative of the United States.
Thank you for the floor.
We continue to have serious legal concerns with some of the remaining paragraphs.
We've repeatedly stated that we cannot accept them as currently drafted because they either misstate IHL or exceed the scope of the CCW and the mandate of the group.
Resolving the concerns related to these paragraphs is also essential to reaching consensus this week.
For paragraph 27, our key concern is with the final clause.
IHLs application, including the CCW protocols, can depend on the technology used.
For example, some rules apply to incendiary weapons while some rules apply to mines.
However, the rules governing attacks apply regardless of the technological means used to conduct the attack.
This concept exists in Article 59 of Additional Protocol one and can be easily reflected the paragraph would read, and includes rules that are applicable regardless of the technology used to conduct the attack.
For paragraph 31, we have a problem with the word which.
This issue can be addressed if the report uses if they are, as done in the 2019 and 2023 reports.
For example, laws must not be used if they are of a nature to cause suprofus injury, et cetera For paragraph 32, as we have consistently stated, this language is unclear and duplicative of paragraph 35.
Although we can work with the concepts of limiting and anticipating, we do not support creating a new prohibition that is already covered by existing prohibitions.
For paragraph 33, the additional suggestion proposal does not resolve our legal concerns with this paragraph.
The phrase protected as such should modify civilians rather than civilian objects, which are, by definition, protected under IHL.
The phrase could be revised to say civilian objects and individual civilians protected under IHL to address this concern.
In addition, we continue to have legal concerns with the formulation object of attack by laws.
Laws are not persons, so they do not have an object of attack.
To address this issue, we propose revising the last phrase to read the object of an attack by means of laws to clarify that the weapon system is the means of attack.
The paragraph would read, It is prohibited to make the civilian population as such, as well as civilian objects and individual civilians protected under IHL, the object of an attack by means of laws.
For these and other paragraphs, we previously proposed a number of different alternatives that we believe would convey the legal obligations correctly.
We remain willing to work with you and other delegations on these and other paragraphs.
For paragraph 39, we do not agree with the must in the SAP, and we do not agree with across the life cycle of laws.
For example, the obligation in the 1977 Additional Protocol one me specific stages of the life cycle.
We also think it is critical to add the word new, which is used in the 1977 Additional Protocol one.
Legal reviews are conducted of new weapons.
Rather than trying to renegotiate these issues, we believe a better approach would be to use the language from guiding principle E, which was carefully negotiated.
That would save the group time.
We also continue to have concerns that we previously expressed with regard to subparagraph B.
We propose either to delete the paragraph or to replace it with the following.
Modifications to a laws whose employment has been determined not to be prohibited in all circumstances, that could render the use of the laws prohibited in all circumstances or that could substantially change the circumstances in which it may lawfully be used should also be reviewed, period.
Regarding paragraph 40, we continue to believe that Chabot needs work as there should not be an abstract requirement to understand the weapon system.
No such norm exists under IHL.
Like other paragraphs I've highlighted, this one needs a connection to existing IHL rules.
Additionally, the term across the life cycle of laws does not make sense because testing and evaluation is a part of the life cycle of laws.
This terminology also anthropomorphizes laws, which we think should be avoided.
In a spirit of compromise, we are open to working with some of these terms if we can explain how they relate to existing rules of IHL like distinction, proportionality, and precautions and attack.
We recommend revising, so the text states, States should, as appropriate, ensure that the capabilities and limitations of laws in their expected use and their anticipated effects in different circumstances can be understood and used in compliance with the IHL rules on attack by conducting testing and evaluation, including simulations to assess the reliability of laws, including their identification, selection, and engagement functions.
The concepts in subparagraph B regarding training, guidance, and instructions are already addressed in other paragraphs.
While subparagraph C raises serious concerns we have tried to include elements of C in our proposal above.
Regarding paragraphs 41 and 42, these paragraphs are not tied to the use of loss.
Additionally, they are not related to IHL.
It is important to connect these paragraphs to the mandate of the CCW and this group.
We also note that the term prevent is duplicative of mitigate in this context and the phrase as much as possible seems unwarranted given the early stage of state practice in this area.
For paragraph 41, we recommend adding the phrase and other technological components after artificial intelligence models to make this paragraph more flexible as the kinds of technology used will continue to change.
To address this and the other issues above, Paragraph 41 could read, States, states should implement measures to mitigate unintended bias in laws, including bias in datasets, artificial intelligence models, and other technological components, that may increase the risk that the laws will identify, select, and engage civilians or civilian objects.
Paragraph 42 could read, States should implement measures to mitigate automation bias in the use of laws that may increase the risk of harm to civilians and civilian objects.
However, our overarching concern is that paragraph 42 is best placed as one of the measures addressing responsibility and accountability.
Automation bias is fundamentally a challenge for personnel and so should be tied to other similar measures.
For example, it could be incorporated into the SEO for paragraph 45 to read, to promote human responsibility and accountability in the use of laws and to mitigate automation bias and the risk of unattended engagements, et cetera Regarding paragraph 43, and accountable should be deleted from the first sentence and responsibility should be replaced with accountability.
This more clearly aligns with the correct understanding of those terms.
We also propose to delete the second sentence of paragraph 43 as it is unclear.
Altogether, the paragraph would read, States parties to armed conflict, individuals remain at all times responsible under applicable international law for their decisions with regard to laws since accountability cannot be transferred to machines.
Regarding paragraph 44, the phrase, the means or methods of warfare involving the should be replaced with there.
Regarding paragraph 46, this is not a correct statement of existing obligations.
Moreover, states are not and should not be obliged to prevent and suppress violations by other states or parties to a conflict.
A revised paragraph 46 would read, States should take all appropriate steps to prevent and suppress violations of applicable international law, including IHL relating to laws under their jurisdiction.
Paragraph 47 misstates international law by characterizing a good implementation practice as an obligation.
We propose replacing must with should, and we support the change to there.
Regarding Paragraph 48, we believe the chair has moved this paragraph here from its previous position as paragraph 39 to help clarify that the formulated elements or all of the preceding paragraphs.
However, this new placement does not fix the fundamental problem that is unclear what states are being asked to do.
The phrase should be taken into account also dilutes the legal obligations listed above.
As a result, we propose cutting this paragraph in its entirety.
I thank you, Mr.
Chair.
I thank the distinguished representative of the United States for making comments.
We will carefully study them.
Let me just make one remark that we have some difficulty in accepting, if we talk about applicable international law to make a must shot.
I think the application of that is a must.
Next, on my list is Turkey to be followed by the United Arab Emirates.
But let me first pass it to the distinguished representative of Turkey.
Thank you, Chair.
Good morning, colleagues.
I'll go right into paragraph 34.
Here we reiterate a concern raised in earlier consultations.
Unlike most of the surrounding text, this paragraph uses shall, which does not restate an existing role of IHL, but creates a freestanding obligation touching on production, acquisition, stockpiling, and transfer.
This has direct implications for national defense planning, procurement, and industrial policy.
We do not believe it is appropriate to legislate on these matters through a non binding set of elements.
We ask that this paragraph be deleted.
This is not merely a drafting preference.
IHL governs the use of weapons in armed conflict.
It does not, as a rule, impose blanket prohibitions on manufacture, acquisition, stockpiling, or transfer outside of specific treaty regimes negotiated for that purpose, such as those governing chemical weapons, biological weapons, anti personsonal mines or cluster munitions.
A category Sin.
Express what shall sits uneasily with the technical IHL restating character that other delegations have themselves attributed to the text.
As once again that the paragraph be deleted.
Paragraph 39, on 39 A.
On paragraph 39 A, Incorporation of language drawn from Article 36 of Additional Protocol one, specifically study, development, acquisition or adoption.
As this room, my country is not party to that protocol.
We do not believe language specific to one treaty regime should be imported into a document intended to reflect common understanding across all states, including non parties.
We ask that this phrase be deleted and that must in the Chapo be replaced what should.
On 39 B, we reiterate our earlier request that the word ensure be replaced with endeavor to ensure or seek to ensure.
Paragraph 40 C.
We are concerned by the removal of the qualifier adequately before predictable and reliable.
Absolute predictability is not a realistic technical standard for learning systems.
This formulation, read literally, could be understood as requiring a guarantee that cannot be given for any complex system, human operated or otherwise.
We ask that adequately or an equivalent qualifier such as sufficiently or to the extent feasible be restored in the text.
Other delegations, including Norway, France, and Singapore made similar proposals during informal consultations.
We're ready to work on the basis of these formulations.
Paragraphs 41 and 42.
My country has no difficulty with these paragraphs as drafted.
The recommendatory language and the qualifier as much as possible appropriately reflect that bias mitigation is a continuing technical effort rather than an absolute guarantee.
We would only note for the record that measures to prevent and mitigate bias should not be read to discourage the development of AI models resilient to communications denial or jamming environments.
Paragraph 43, we support this paragraph as drafted.
The principle that responsibility cannot be transferred to machines and remains with states, parties to conflict and individuals is well established.
Now moving to paragraphs 44, 46, and 47, which are the crux of the matter for my delegation.
On these paragraphs, we maintain the graduated approach expressed in earlier sessions.
Our preference remains for should rather than must throughout.
On paragraph 44 in particular on individual accountability, the principle of individual criminal responsibility for international crimes is, of course, well established in customary international law.
What remains unsettled in our view is its application to engagements in which a system identifies, selects and engages a target without human intervention.
In such cases, establishing which individual bears responsibility on what factual basis raises questions that this group's mandate has not resolved and that we would not wish this text to prejudge by using unconditional language.
We believe this is best reflected through should with any residual concerns addressed in national statements.
On that basis, we would ask that must ensure in paragraph 44 be replaced with should ensure to avoid the impression of a categorical result obligation in an area where the precise attribution of individual criminal responsibility for engagements carried out without human intervention in target identification and engagement remains, in our view, an evolving and not fully settled area of law.
Should the verb must be retained in this paragraph, we could accept must uphold as an intermediate formulation on the understanding that the existing reference to their obligations under IHL is preserved.
Turning to paragraph 46, which refers to violations of international law, including IHL, without qualification.
We would ask that applicable be inserted before both international law and IHL.
We would further ask that on territory under their jurisdiction be replaced with the words on their territory.
This is not to question the well settled principle that a state bears responsibility for the conduct of its own organs, including where they act beyond their authority or contrary to instructions.
That principle is not in dispute.
Our concern is narrower.
The jurisdiction based formulation, read through the doctrine of effective control, developed in a different body of case law, risks being understood as extending a state's preventive obligations beyond its own territory in a manner this text was not designed to address.
On paragraph 47, the requirement of effective internal mechanisms goes beyond the restatement of the underlying obligation to investigate and report and begins to prescribe institutional form.
We would ask, as with the preceding paragraphs, for must to become should and for applicable to qualify IHL.
As an intermediate formulation, we could support must provide for appropriate internal mechanisms, which softens the institutional implication while retaining mandatory language.
If must is retained in these paragraphs, we ask that applicable be added before references to IHL in paragraphs 46 and 47, consistent with our position elsewhere in the text.
We will reflect our understanding of the scope of these obligations in our closing statement.
The caveats in paragraphs 46 and 47 are necessary if we are to be in a position to join consensus.
For paragraph 44, must uphold instead of must ensure could be acceptable for my delegation.
Paragraph 48.
Here we support chair's decision to retain should as the operative word, and we cannot accept changing this word.
As we noted at the outset, we're also concerned by the addition of and other possible measures.
This phrase leaves the scope of expected measures undefined and open ended and could be read as inviting the future inclusion of measures not agreed by consensus.
We ask for its deletion.
Alternatively, we could support qualifying the phrase as other possible measures as may be agreed by consensus, tying any future additions to the group's established working method.
Last but not least, my delegation wishes to formally propose the insertion of a clause either at the outset of subsection A or in its chapot along the following lines.
These elements are without prejudice to the nature of any possible future instrument and do not in and of themselves create, alter, or expand rights and obligations of states under international law.
This language tracks closely with the qualifiers already found in paragraph 48 of the report.
We believe such a clause would go a long way toward resolving many of the concerns we have raised regarding the use of masks throughout the text.
We understand from informal exchanges that the chair's own explanation of this text, that it restates existing obligations and creates nothing new, is not controversial among delegations.
If that is so, there should be no obstacle to reflecting it expressly, which would also reduce the need for delegations, including our own to rely on lengthy national statements to the same effect.
We would welcome the opportunity to discuss the placement and precise wording of such a clause with interested delegations.
Thank you, Chair.
Let me thank the distinguished representative of Turkey.
Just two remarks from my side.
I think we have to be a bit careful not introducing the concept of braces and belt in the text by changing a lot of verbs, referring to remarks by the chair and delegations and then also insert a clause and all to cover the same issue.
I think this is something we have to reflect on.
Then another remark with regard to what was said on establishing individual accountability and possible individual criminal responsibility, that it is difficult with laws, I think exactly for that reason, and I think it was very well explained by our Pakistani colleague, we have introduced terms like traceability, predictability, reliability.
So I would really recommend also to To read this as a set of elements.
It's not just a loose number of faults.
We try to give it some good considerations.
That's from my side as a short reflection, so far.
Next on my list is the United Arab Emirates to be followed by the representative of IHLYI.
But first, the distinguished representative of the United Arab Emirates Y.
Thank you, Mr.
Chair.
The United Arab Emirates would like to submit the following observation regarding the paragraphs under discussion.
Regarding paragraph 39, we approve the general approach of this paragraph, namely the importance of preserving the two expressions as appropriate and in accordance with their obligations under international law.
These are very important and we prefer to keep them in the text regarding paragraph 39 A.
We approve the legal revisions or reviews during the examination of laws and developing them, adopting them according to the obligations applicable to countries as per IHL.
We approve the a voluntary exchange of practices between countries while preserving guarantees pertaining to national security and commercial considerations protected by intellectual property considerations.
For us and the UAA, these restrictions are very important in order to preserve the voluntary nature of the exchange of practices and to protect the legitimate interest pertaining to national security and sensitive technological information.
Regarding 39 B, we can approve the current formulation, and there is an importance of adding amendments that have a substantial impact on determining the objectives, choosing them or engaging with them or their anticipated effects.
For us, the substantial impact as a criteria is important, should be kept.
It secures the revision of amendments that can have a substantial impact on the performance of the system, its impact, rather than reviewing in an automatic way for a minor amendment.
Regarding paragraph 40, we approve the general objective, namely to have a sufficient understanding of the capacities of laws, their impact, and their anticipated impact in different circumstances regarding Paragraph 48, we approve the conducting of testing and evaluation, namely when it comes to simulations.
However, we think it's important to have a clear link between the testing and the operational conditions under which the system is used.
In the case of a need of further clarification of the formulation, we can support the following, including simulations designed to assess their performance under anticipated operational conditions under anticipated operational conditions.
Regarding 40 B, we approve the current formulation when it comes to ensuring appropriate guidance, training and instruction for those in the responsible chain of command and control of flows and we think that the appropriate guidance and understanding of the capacities of the system are important factors that enable human beings to take informed decisions.
Mr.
Chair, regarding paragraph 40 C, we approve the objective to ensure an appropriate level of predictability and reliability in terms of laws.
These two factors are very important in order to determine whether this system can be used in line with IHL.
However, we consider that the principles of explainability, traceability, and predictability are still principles that are discussed inside our group and there is no technical or legal definition that is unified and agreed upon about these principles.
That's why we think that the We should refrain from explaining this concept as absolute technical requirements or unified requirement.
For us, the main criteria is to have a system that is operated in a way that is well understood predictable manner so that those who are in charge can take informed decisions and secure observance and and compliance with IHL applicable IL.
Regarding paragraph 47, we approve the presence of efficient national mechanisms for investigation and to take measures in cases where the use of laws may involve violations of IHL.
However, The paragraph uses, states must ensure.
However, we disagree.
We think that this commitment should be clearly linked to applicable international law, and we proposed the following in accordance with their obligations under international law.
Order to clarify that this paragraph aims at implementing existing legal obligations and do not create a new commitment or a separate obligations that is wider when it comes to investigation and reporting.
Regarding paragraph 48, Mr.
Chair, we understand the objective.
That is to take into account the formulated elements when it comes to the design, development, and use of flows.
However, we require more clarification regarding other possible measures.
This is an open expression.
It is not limited and could lead to uncertainty when it comes to other possible measures that countries should take into account.
That's why we prefer either to keep the text clearly linked to the elements that have been agreed upon inside the group or in case of mentioning additional measures to clarify the following as appropriate and consistent with states obligations under international law.
Mr.
Chair, this should prevent us from creating an open expectation about measures that have not been defined or agreed upon clearly.
We want a clear text that is based on legal division, testing, evaluation, appropriate training, and a sufficient amount of predictability and reliability and to have efficient accountability mechanisms at the national level.
The text should clearly clarify the existing legal obligations, the technical guarantees, best practices that are recommended while preserving considerations pertaining to national security and to provide sufficient flexibility in order to take into account the various technological and operational contexts.
Thank you, Mr.
Chair.
Let me thank the distinguished representative of the United Arab Emirates for commenting on the text and making proposals.
Just allow me one remark with regard to paragraph 48, where we have to formulate elements and other possible measures should be taken into account, blady, blah, blah.
That comes from our mandate.
In the context of the objectives and purposes of the convention, the group is to further consider and formulate by consensus a set of elements of an instrument without prejudging its nature and other possible measures to address emerging technologies in the area of lethal autonomous weapons systems.
So there it comes from.
There are already a few caveats in that, of course.
It's not that it's an open door, it's about what we agree.
And then there's still the issue of the nature of the elements and other possible measures.
But let's see.
If that's not clear, we might have to rephrase it, but it's not something just dropping out of the sky.
Next on my list is the distinguished representative of IHLYI to be followed by the ICOC and New Zealand.
But first, the distinguished representative of the IHLI.
You have the floor.
Thank you, Sir.
We'll be following the discussion and has some points we believe are important to add regarding paragraphs 35 to 39.
IHLYI supports the current formation of paragraph 35.
The microphone, please for the speaker.
Something going wrong with your microphone.
Could you try again? Hello? Perfect.
Thank you.
IHL supports the current formulation of paragraph 35, in particular, the reference to applicable international law, in particular, IHL.
We consider this formulation important to ensure that the use of laws remain grounded in the applicable legal framework, while reaffirming that compliance with IHL, including the principles and requirements of distinction, proportionality and precautions in attack must guide their use.
We also welcome the recognition in paragraph 36 that human judgment and control are necessary to uphold compliance with international law, in particular IHL.
We appreciate that the text does not equate this requirement with constant direct human control.
We also express our appreciation for Ireland's proposal outlined in this paragraph.
We support the measures set out in paragraph 38 B, particularly the ability to limit the effect of loss by restricting the types of targets, duration, geographical scope and scale of operations.
However, we believe we should retain the original wording of Article 38B, which includes the number of engagement of loss.
Such parameters are important practical safeguards for ensuring meaningful human judgment and control and for reducing the risk of intended engagements.
Concerning paragraph 38 C, we endorse the removal of the phrase while active in an operational context.
IHLYI also welcomes paragraph 39, particularly its recognition of the need for legal reviews as appropriate across the life cycle of laws and in accordance with states obligations under international law.
In this context, we would like to reiterate an issue that IOYI has consistently raised in its previous interventions concerning the role and responsibilities of private sectors of private sector actors across the life cycle of loss.
We recognize and appreciate that the current text already addresses the development, use, testing, evaluation, responsibility and accountability of loss across the life cycle.
We therefore do not seek to introduce a fundamental change to the structure or substance of the current text.
Rather, we believe that the existing life cycle approach provides an important opportunity to clarify how the responsibilities and relevant actors beyond traditional defense institutions can be addressed in practice.
The development of autonomous and AI enabled military technologies increasingly involves technology companies, software developers, data providers, research institutions, and other private sector actors.
Their contributions may affect the design, capabilities, testing, reliability, and eventual use of these systems.
As we have emphasized in our previous interventions, we encourage the group to consider how states can ensure appropriate due diligence, risk assessment, transparency, and oversight in relation to relevant private sector directors involved in these processes.
Our intention is not to introduce new or separate obligations for private companies through this text, but rather to ensure that the existing principles of responsibility, accountability, and life cycle oversight are not undermined by increasingly complex technological supply chains.
We believe that a clearer understanding of these roles will strengthen the implementation of existing elements without requiring a fundamental change to the text.
Lastly, we wish to express our gratitude for the efforts of the chair and we would like to add that a copy of this statement will be shared with the Secretariat.
Thank you, Chair.
Let me thank the distinguished representative of IHLYI.
Next on my list is the ICSC to be followed by New Zealand, Brazil, Ireland, and the Philippines.
But let me first pass the floor to distinguished representative of the ICSC.
You have the floor.
Thank you, Chair.
We welcome the paragraphs in this section.
In particular, we support the inclusion of the paragraph on legal reviews, as well as the language in your additional suggestions, which aligns with existing IHL language in Article 36 of Additional Protocol one.
We also echo the remarks made by Pakistan and the UAE on the importance of retaining Paragraph 39 B, which helps ensure that legal reviews continue to serve the purpose for which they were designed, including in addressing the challenges posed by new technologies.
Moreover, we welcome paragraph 48.
However, we would advocate for retaining the concept of realistic simulations to capture the need for these to match as closely as possible the ultimate environment of operation.
That said, we've also listened to the proposal by the distinguished representative of the Russian Federation with interest and wonder if the following formulation could be an acceptable alternative suggestion.
It could read as simulations that, to the extent feasible, are representative of and can assess the loss performance within the expected operational environment.
We also welcome the inclusion of predictability, reliability, traceability, and explainability in paragraph 40 C.
As the distinguished representative of Pakistan has noted, these are important bridging concepts for ensuring compliance with IHL, including by enabling users to anticipate and limit the effects of the weapon system.
We also welcome your additional changes to the paragraph chair, which improve its clarity and precision and clarify that all four bridging concepts relate to both the systems functioning and its effects.
More broadly, we consider that the objective of the group's work is not only to restate existing IHR rules, but to also clarify how they apply in the context of autonomous weapons and to articulate additional specific measures that respond to the particular challenges raised by such systems.
This approach aligns with the precedent set in CCW amended Protocol two, where Article 3 restates general IHL rules and principles applicable to all weapons, and Articles four to seven provide additional specific prohibitions and restrictions on the use of mines, booby traps, and other devices adapted to the specific characteristics of these devices.
I thank you, Chair.
And I thank the distinguished representative of the ICOC for commenting on the text and making proposals.
Next on my list is New Zealand to be followed by Brazil, but let me first pass the floor to Distinguished representative of New Zealand.
So you have the floor.
Thank you, Chair.
I feel a little bit boring because we can largely live with the text on these paragraphs as a set of elements as you've put up on screen, rather than attempting to element at this stage.
It is a little difficult to follow, let alone react to all of the proposals that have been made thus far this morning.
I'm not going to attempt to do that at this moment.
We'll need to try and digest some of those and come back to you later.
But just to note, we would support comments from Pakistan and ICRC about the importance of some of those various paragraphs.
I just wanted to note two issues that we had with your text and that's in paragraph 39 B.
We continue to have concerns about the undefined threshold of substantial impact in that paragraph.
A review may itself be necessary to determine whether a modification has a substantial impact.
Our preference would therefore be to delete that qualifier or to refer instead to modifications that may affect the relevant functions or anticipated effects.
We could, for example, use something like material impact.
On paragraph 48, we would continue to prefer must over should, but we could probably live with should.
Thank you, Chair.
Let me thank the distinguished representative of New Zealand.
Next on my list is Brazil to be followed by Ireland, but let me first pass the floor to distinguished representative of Brazil.
Thank you, Mr.
Chair, and good morning.
Turning to the paragraphs that we have not addressed yet on paragraph 39, my delegation like those of Pakistan and the ICRC believe that this is one of the essential elements of the text and we believe that is particularly true with regards to 39 B, which addresses how the obligation of legal reviews applies specifically to laws, which are systems that have a nature of continuous evolution and development.
We appreciate the delegation of the US for proposing alternative language for this paragraph.
At first sight, that formulation might help.
Circumvent the long debate we have had on the issue of what constitutes a substantial impact on the operational function of laws, and while we remain attached to this language, we will study the proposal by the US carefully.
Turning now to paragraph 40, also like Pakistan and the CRC, we believe that these elements in this paragraph, in particular, the concepts of predictability and reliability, explainability and traceability are also central to our discussions on regulation of laws.
We believe that the concepts of predictability and reliability stem directly from the requirement of IHL on the need to anticipate and limit effects of the use of weapons, and the concepts of explainability and traceability are fundamental to ensure accountability in their use.
On paragraph 40 A, we take note of the changes proposed by you, Chair.
While we would have preferred to retain a reference to the need to ensure that simulations are realistic, we can accept the formulation as it is.
However, we would caution against any further changes to the language that might suggest that laws might be tested in real operational contexts.
Um, subparagraph 40 C also I'm sorry, ready to address that.
On paragraph 47, we also attach particular importance to this language.
We believe that the mechanisms for investigation and reporting of potential violations is nothing new.
It is aligned with longstanding practice for disarmament and arms control mechanisms, and it's present in many other such instruments.
As to paragraph 48, our delegation also believes that this is extremely important.
We take note of the arguments presented regarding the relocation of this paragraph, and we highlight the fact that taking these elements in this text into consideration is fundamental to ensure that development of laws evolves in alignment with IHL.
We take note of the concerns raised by the delegation of Turkey with regard to the reference to other possible measures, and we might consider those comments favorably.
Thank you, Chair.
Let me thank the distinguished representative from Brazil.
Next on my list is Ireland to be followed by the Philippines.
But let me first pass the floor to distinguished representative of Ireland.
Thanks very much, Chair.
Regarding the remaining paragraphs of the rolling text, we support 39 Chabot in the sub paragraphs as you have drafted.
The Chabot has the as appropriate caveat which should address some of the existing concerns that we have heard.
As a general point, as is the case elsewhere in the text regarding the debate over should versus must, we do not think we will find agreement on this here this week, and we think that this can and should be dealt with in future negotiations.
We see Meritan retaining the word realistic.
Purely virtual simulations would not be sufficient for robust testing and evaluation necessary to have a reasonable level of certainty of the system's performance in various dynamic scenarios.
So we would prefer to retain it and we welcome the addition of can assess their performance.
But we also thank the ICRC for their textual suggestion just now, which at first reading, we think can help bring some more clarity.
On 39 B, we heard ideas such as replacing substantive with material as New Zealand has just suggested, which is something I think we can look at positively.
On 40 C, we very much welcome that predictability, reliability, traceability, and explainability have been retained in these elements.
Any text containing elements of an instrument on issues related to AI and autonomy need to include references to these concepts that are well known and understood across multiple disciplines and we thank Pakistan, Brazil, and the ICRC for providing a strong case for why these concepts are essential here.
We can go with the changes that you have made, which do bring some clarity.
For us, what's vital is that the concepts are referenced so that our successors can carry them forward.
Their absence would undermine the effectiveness of any future instrument.
Regarding paragraph 48, we're okay with its placement here as it's cross cutting across the entirety of the elements, and we support it as drafted as an essential part that should be retained.
Chair, we heard this morning some interventions that have gone through these paragraphs and reduced nearly each one to what is the bare minimum as required by IHL.
We want to reiterate that IHL is a floor, not a ceiling, and that the purpose of the CCW and as such, this group is to implement, operationalize, and progressively develop the fundamental rules of IHL that prohibit or restrict the use of conventional weapons deemed excessively injurious or to have indiscriminate effects.
Each CCW protocol does this in a different way.
Amended Protocol two in particular contains concepts and obligations, including technical design specifications that clearly go beyond baseline IHL.
It contains engineering parameters that general IHL does not clearly prescribe.
Our mandate includes taking into account existing protocols, so we should keep that in mind as we develop these elements.
The clear and unique challenges posed by laws require this group to develop and strengthen IHL in line with the mandate and the purpose of the CCW and in line with its existing protocols.
We also note that in many of the paragraphs of the rolling text that have drawn concern regarding going beyond IHL, the operative word is should, which clearly signifies that it is not suggesting obligations to go beyond IHL.
I think.
Let me thank the distinguished representative of Ireland for stating the positions of his delegation and commenting on proposals made earlier this morning.
Next on my list is the Philippines to be followed by France, India, the ROK and Armenia.
But let me first pass the floor to distinguished representative of the Philippines.
You have the floor.
Thank you, Chair.
As it is the first time our delegation takes the floor, I would like to thank and recognize you and your team's efforts to accommodate the positions of each and every delegation in this group and your hard work in trying to arrive at a consensus by the end of this week.
Mr.
Chair, allow me to use this opportunity to express that my delegation supports the June five version of the ruling text as we deem it to be a balanced and well structured version of the said document.
Nonetheless, and in the spirit of compromise and collaboration, we will offer our comments on the text incorporating the additional suggestions by the Chair.
Before diving into our textual comments, my delegation would like to underscore the dichotomy that is prevailing in this body, which we view as a fundamental factor that contributes to the challenge of arriving at a consensus.
After listening closely to the interventions for the past two days and in the previous informal consultations, there remains two camps in this body.
Those who deem that IHL is the sole instrument that governs armed conflicts, hence should be the only instrument that should govern the use of laws and that should be satisfied by the measures proposed to govern laws.
On the other hand, there are those who deem that the complexity of autonomous weapons systems warrants further guardrails that might go beyond the existing IHO.
Hence, additional measures should be observed and complied with.
Our delegation hopes that the gap between these two schools of thought will be narrowed as we approach the end of this week.
Mr.
Chair, with regard to your additional suggestions on paragraphs 39 to 48, my delegation supports the elaboration of the conduct of legal reviews in paragraph 39 A.
On paragraph 48, we support the inclusion of conducting testing and evaluation of laws and views the process to be within the life cycle of laws.
On your additional suggestions in other sub paragraphs in the name of compromise and the goal of arriving at a consensus, we accept your additional suggestions.
With regard to the comment on paragraph 47, wherein there is an implication that it is a form of intervention on internal state affairs, Our delegation argues that the spirit of such obligatory measures has also been adopted in the 2019 report of the GGE and thus supports the retention of the said paragraph.
This is all for now from our delegation.
Thank you, Chair.
Let me thank the distinguished representative of the Philippines for commenting on the text.
Next on my list is France will be followed by India.
But let me first pass the floor to distinguished representative of France.
President.
Thank you, Chair.
We'd like to thank you for your suggestions.
My delegation believes that the text as a whole is a good basis upon which we could deliver our mandate.
That being said, a number of concepts and sections are still under debate.
We would like to make a number of suggestions.
We already presented these during informal consultations, but in our view, they could help clarify some of the paragraphs.
Regarding paragraph 34, we understand that it is through paragraph 34 or through this paragraph 34, rather that these prohibitions apply to prohibited systems under paragraph 34, that is to say, those which cannot be used in compliance with international humanitarian law.
Scope of this measure and its implementation will, however, require further technical discussions subsequently in order to ensure that we do not unduly hamper research, to ensure that it would be possible to develop certain systems not for the purpose of using them, but rather for developing countermeasures and to untangle issues of accountability in the event of a transfer of a system.
Therefore, for now, we would like to suggest the following wording for paragraph 34.
States shall refrain from manufacturing, otherwise acquiring, stockpiling, or transferring loads which are incapable of being used in compliance with IHL clasp of being used.
Matt paragraph.
Moving on to paragraph 40 A.
We welcome the reintroduction of the concepts of predictability, reliability, traceability, and explainability.
These are important notions for ensuring respect for international humanitarian law requirements.
However, it is also important to bear in mind that we're addressing specific weapons here that will integrate AI algorithms and automatic machine learning and that total explainability may not always be possible or necessary.
That's why we think it would be wise to add a qualifier such as adequately or sufficiently, which is taken from the 2023 report.
We would therefore like to suggest the following rewording.
L at the same paragraph, but with the addition of The functioning and effects are adequately predictable, explainable, and traceable the addition of the term adequately before predictable, explainable, and traceable in this paragraph is what we're suggesting.
Moving on to paragraph 46.
The phrase, take all appropriate steps to prevent and suppress violations comes specifically from international humanitarian law.
Obligations under the rest of international law are worded differently.
We therefore believe it would be preferable to draw a distinction between wording taken directly from IHL on the one hand and those coming from international law on the other.
One possible reformulation could be as follows, but of course, we are open to other suggestions, and I read, States, ensure that persons and legal entities acting under their authority or control comply with their obligations under international law That's the addition that we're suggesting there.
The graph and the rest of the paragraph would remain unchanged.
Just take all appropriate steps to prevent and suppress violations of IHL relating to laws by persons and persons on territory on their jurisdiction.
Monsieur President Chairman, I thank you.
Let me thank distinguished French ambassador for commenting on the text and making proposals.
Next on my list is India to be followed by the Republic of Korea, Armenia, Bangladesh, Israel and the Russian Federation.
But let me first pass the floor to the distinguished representative of India.
You have the floor.
Thank you, Mr.
Chairperson, and good morning, colleagues.
Mr.
Chair, we want to appreciate you and your team's work, and we also want to thank you for convening us in formal and informal sessions in steering our work ahead.
My delegation has a few preliminary observations to offer on the paragraphs that we're discussing, and we will pay heed to your plea, which is that we will not introduce a lot of proposals and deletions given the paucity of time, but we will indicate what is of necessary importance to my delegation.
Mr.
Chair, on paragraph 34, which refers to the scope of prohibitions, we have been conveying our reservations on the expansion of prohibitions from use to cover manufacturing, acquisition, stockpiling, and transferring, particularly given the manner in which this formulation is, you know, is made, we want to underline that IHL prohibitions generally cover use, and we want this instrument to be consistent with that benchmark.
But we are willing to work in addressing this paragraph.
In the previous draft of the text, this was earlier balanced by another sentence on permitting R&D for countermeasures for training and other purposes.
And because of these two sentences, there was a balance, uh in the measure, uh, uh which is given in paragraph 34.
So our suggestion would be that we work on that basis.
And we also want to underline that paragraph 48 already acknowledges that the formulated elements and measures, should take into account the design, development, and use of laws.
That aspect to a large extent covers the concern which we are trying to put in using paragraph 34.
So our suggestion would be that we could also avoid repetition, given that paragraph 48 covers a large, part of the concerned that is stated in paragraph 34.
Mr.
Chairperson, on legal reviews, in paragraph 39, we want to suggest a minor editorial reordering of the Chapot dealing with the legal reviews.
The Chabot part should start as follows, which is in accordance with their existing obligations under applicable international law, states must across the life cycle of laws, and then we can continue with the subparagraphs.
Our view is that this will bring more clarity to it and clearly state the differential obligations that states have as per the applicable international law.
I also want to state that paragraph 39 A makes a reference to voluntary exchange of relevant best practices, and this understanding has been agreed by all states under the GGO laws.
We would also support taking out the second sentence of paragraph 39 A as 39 B, which applies to all states, which is that these exchange of relevant best practices should be encouraged, bearing in mind the national security considerations and commercial restrictions on proprietary information.
Mr.
Chair, we can support the Russian edit on paragraph 48, which seeks to embed the simulation in realistic environment, and we also are happy to support the edit proposed by the ICRC on ensuring that these weapons systems are better tested in the realistic environment to better understand and improve the weapon performance.
Uh, on paragraph 40 C, my delegation would just like to add that we are in the spirit of flexibility, willing to work on terms predictable and reliable.
I think, a point was made that there is a need to ensure that human judgment and control, is able to ensure that weapon performance is consistent with the anticipated effects of the attack.
So we're willing to work on these two terms, although this would require some reformulation.
However, we at this stage have reservations on the term explainable and traceable in the present form.
I want to note that these are not IHL specific terms and they lack a common interpretation, and operationally, we do not really know what they would entail in terms of weapon design and development.
We have reservations on the term explainable and traceable, but we are willing to work on the terms predictable and reliable.
Of course, we've noted edits by the Turkish delegation and others on using the right adverb before predictable and reliable, such as adequate level of predictability and reliability, we're willing to work on those formulations.
And, of course, Mr.
Chairperson, my delegation can also support paragraphs 41 and 42, as you've suggested, which contain very important concepts of biases.
And in paragraph 41 and 42, we cover three different kinds of biases, algorithmic biases, data biases, and automation biases that come from human machine interaction.
We feel that these paragraphs are in right, formulations.
Now moving on to paragraph 47, Mr.
Chairperson, we can accept paragraph 47 as a general principle, and we understand and appreciate the principle that is behind this provision.
However, we want to note that this does not come from existing IHL.
As one delegation pointed out, that we are converting a best practice, uh, you know, into an obligation over here.
We are willing to work on that aspect too.
Uh, we are open to retaining the term internal mechanisms.
However, we would like to keep these internal mechanisms, whatever they may be open ended without pointing to what these internal mechanisms may be for states, as states may clearly have different internal structures on dealing with, uh, uh, on dealing with, uh, you know, responding to incidents involving, uh, the use of laws.
And lastly, Mr.
Chair, this is a concern we've raised before, and we want to know if other delegations have any particular views with this respect, which is on the clear risk of use of laws by non state armed groups and by terrorists.
And there are three points I want to make.
First is that unlike states, non state actors, including terrorists do not have to comply with IIHL.
And second is that concerns regarding laws, particularly that come from autonomous targeting apply more in the case of terrorist and non state armed groups.
Such actors have no intention to limit civilian casualties.
In fact, by design, they intend to target civilians.
And third, the group has already acknowledged these risks.
The guiding principle F talks about the risk of acquisition of laws by terrorist groups as well as the risk of proliferation.
This is further complemented by paragraphs 23 A of the 2019 report that further emphasizes such concerns.
So overall, our suggestion would be that states should undertake obligations committing not to transfer laws to terrorist and terrorist groups and on state armed groups.
Additionally, states should further ensure proper tracing and record keeping of the weapon system across the weapon life cycle to prevent acquisition and diversion.
Would be happy to share this formulation to you and we'll be happy to work with other delegations in having this enshrined in the text.
With this, Mr.
Chairperson, I conclude my intervention and we have not addressed many proposals that were made, but I'm sure that once these are reflected on to the next text that you will propose, we will come back and suggest our edits.
Thank you.
Let me thank the distinguished representative of India for commenting on the text and making proposals.
Indeed, I also invite other delegations to reflect on the proposal just made, also in the light of our mandates for the group.
Next on my list is the Republic of Korea to be followed by Armenia and Bangladesh.
But let me first pass the forward to the distinguished representative of the DK.
Thank you, Chair.
My delegation is once again ready to join the consensus on the rest of the text that you have proposed with the only exception of paragraph 34.
On paragraph 34, let me reiterate our position.
My delegation does not object to the general purpose of this paragraph.
However, as we have repeatedly pointed out, for this important paragraph to work in practice, we need the exception clause that would allow production in small quantities for research and development, training, and the development of countermeasures.
As far as I understand, the anti Psonal minban Convention from which the wording of this paragraph was borrowed contains such exception clause.
Without similar exception clause, the paragraph is stoo definitive.
Therefore, during the informal consultation, I just suggested replacing the verb Cheryl with may.
In the speed of compromise, however, my delegation could consider should.
We also suggest retaining only manufacturing and stockpiling and deleting acquisition and transfer, given the latter two issues require further discussion to address the concerns expressed by our Turkish colleagues and others.
The paragraph could read as states should refrain from manufacturing or stockpiling laws that cannot be used in compliance with IHL.
We can also consider the French suggestion regarding this paragraph, but even in that case, we strongly suggest that we change the verb.
That being said, let me react to the sum of comments expressed in the room in this morning.
I believe that most of the comments we have heard today are intended to improve the text and ensure that the language is legally precise so that states do not encounter legal difficulties as we further deliberate on the text in the future.
In this context, I found the comments made by our US colleagues on various paragraphs helpful.
If we could see those proposed edits on the screen, it would make it easier for delegations to understand their intent.
Maybe not today, but later this week we can work on this proposals.
Also, given the difficulties we are facing in finding language on the 39, we should consider drawing the agreed language contained in guiding principle E as suggested by the US.
We also found the comments made by our Russian colleagues helpful on a number of paragraphs, including the proposed edits to paragraph 48, which you can consider positively.
On paragraph 40 C, we understand that some delegations have concerns about the concepts expressed in this paragraph.
My delegation can be flexible.
Perhaps we could address this concern by moving the subparagraph to a separate paragraph and using more general language such as states are encouraged to or states should seek to ensure as our Turkish delegation suggested.
In this way, we can retain the substance, the concepts that many delegation would like to see in the text while also addressing the concerns of others.
As we are approaching the midpoint of our week long journey, I would like to reiterate that my delegation attaches great importance to our collective work, which has been built over the past three years under the guidance of our chair.
We therefore hope all the delegations show some flexibility so that we can achieve something meaningful by the end of this week and we can build on this working going forward.
Also, I agree with our Irish colleague that chair is not the ceiling, but the floor or the starting point.
At the same time, I do not think we should be overly ambitious to go beyond the ceiling in every paragraph, nor we should aim merely to meet the floor each time.
We should start to strike the right balance between the two.
I think chair.
Let me thank the distinguished representative of the Republic of Korea for sharing the views of this delegation, for commenting on proposals by others and for also sharing some reflections with regard to the work and the moment in time we find ourselves here.
Next on my list is Armenia to be followed by Bangladesh and Israel, but let me first pass the floor to distinguished representative of Armenia.
You have the floor.
Thank you, Mr.
Chair.
Armenia welcomes the attention given to the question of transfer in the discussion of paragraph 34.
We attach particular importance to the responsible transfer of military technologies incorporating autonomous capabilities.
In our view, the existing provision could usefully be complemented to reflect that in considering the transfer of laws, states should, in accordance with their applicable international obligations and national procedures, assess the risk that such systems could be used to commit or facilitate serious violations of international humanitarian law and take appropriate measures to prevent the diversion to unauthorized end users.
We remain flexible as to the precise wording and stand ready to contribute constructively to further consideration of this issue.
Regarding paragraph 39, Armenia attaches importance to retaining a clear reference to legal reviews in paragraph 39.
Effective legal review procedures are an important safeguard for ensuring that the development and use of laws remain consistent with applicable international law.
We remain flexible as to the precise formulation with a view to accommodating different legal obligations and national procedures.
As for paragraph 42 42, we attach particular importance to ensuring the reliability of such systems through appropriate testing and safeguards aimed at mitigating the risks of malfunction and unintended outcomes.
In this regard, we consider the references to predictability, reliability, traceability, and explainability to be useful elements of the text.
We also support retaining the provisions addressing unintended and autonomous bias.
As for paragraphs 43 to 44, Armenia strongly supports the retention of paragraphs 43 and 44 for Armenian responsibility and accountability for decisions concerning the use of laws must remain with human actors and cannot be transferred to machines.
We also attach particular importance to maintaining a well defined and responsible human chain of command and control as an essential component of effective accountability.
Thank you, Mr.
Chairman.
I thank the distinguished representative of Armenia for shame refusal of the delegation and also referring to proposals made by other delegations.
Next on my list is the distinguished representative of Bangladesh to be followed by Israel, the Russian Federation, and Belgium.
But let me first pass the floor now to the distinguished representative of Bangladesh, you have the floor, sir.
Thank you, Mr.
Chair.
My delegation has carefully listened to the viewpoints expressed by the delegations this morning.
We are concerned to hear calls for diluting language in a number of paragraphs that aim to strengthen individual accountability for the use of laws.
My delegation categorically explains why we advocate the use of must instead of should.
Bangladesh approaches the question of individual accountability from a broader concern, the integrity of international legal order and the preservation of individual responsibility within the broader framework of state responsibility.
State responsibility remains fundamental, but for the gravest violations of international law, it is equally essential to identify the individuals responsible and ensure that they can be held accountable.
The international legal order rests on a broader architecture.
The prohibition on the use of force, international humanitarian law, international human rights law, and international criminal law.
These regimes serve different functions, but they are mutually enforcing.
Their common purpose is to place law between the exercise of power and its consequences for the individual.
That architecture ultimately depends on a human being to whom unlawful conduct can be attributed and who can be held responsible.
This is where autonomous weapons raise a fundamental concern.
The issue is not simply whether one unlawful act might escape attribution.
The deeper concern is that if autonomous systems progressively separate human decision makers from the consequences of the use of force, they could weaken the very chain through which international law operates.
If conduct cannot be attributed, responsibility cannot be established.
If responsibility cannot be established, accountability is weakened.
If accountability is weakened, deterrence is diminished, and when deterrence and accountability are diminished, the authority of the legal rule itself is progressively eroded.
This would be more than an accountability gap concerning a particular weapon system.
It could become a structural erosion of the principle that grab violations of international law must have consequences for those responsible.
Bangladesh has therefore supported an explicit reference to individual criminal responsibility in the ruling text.
The principle that responsibility and accountability cannot be transferred to machine must be clear.
Humans, not machines bear individual criminal responsibility for grab violations of international humanitarian law.
This does not mean that commanders, programmers, operators, or manufacturers incur automatic liability.
Established principles governing individual criminal responsibility must continue to apply, but autonomy must never become a legal shield.
Where the elements of an international crime are otherwise satisfied, the autonomous nature of the weapon must not prevent the identification and prosecution of the responsible individual.
Mr.
Chair, our concern therefore is not that laws will themselves cause genocide, crimes against humanity, or other grape crimes.
It is that technological autonomy must not create an environment in which unlawful conduct remains, harm remains, but responsibility becomes increasingly diffused or impossible to establish.
Machines may execute functions, they cannot inherit responsibility.
Technology make complicate attribution.
It cannot eliminate it and autonomy must never become a pathway to impunity.
Thank you, Mr.
Chair.
Let me thank the distinguished representative of Bangladesh for commenting on the text and earlier proposals by delegation and sharing the views of his delegation with regard to those.
Next on my list is distinguished representative of Israel to be followed by Russian Federation, Belgium, Switzerland, Italy, and Sweden.
But let me first pass the floor to distinguished representative of Israel.
Sir, you have the Thank you so much, Sarah, and good morning, colleagues.
Chair, allow me to begin by saying that we will concentrate now on our main comments on these parts of the document.
Regarding paragraph 32, as we previously stated and similarly to other delegations, we have serious concerns with this paragraph.
In our view, it is essential to stick to the language of Article 514C of the additional protocol, the first additional protocol, sorry, to the Geneva Convention and to therefore delete anticipated.
Therefore, the paragraph will read, quote, It is prohibited to use laws in their effects in attack if their effects in attack sorry cannot be limited as required by IHL in the circumstances of the use, end of quote.
We believe that the current text as it stands, introduces new legal standard which is problematic.
IHL requires to make reasonable efforts to assess expected collateral damage based on information available prior to the attack.
It does not, however, establish a separate requirement that the effects of the weapon themselves must be capable of being anticipated as a condition for its lawful use.
Unexpected and anticipated may and often do occur on the battlefield and do not determine otherwise, and to determine otherwise will be unrealistic and counterproductive.
By adding anticipated alongside limited, the paragraph could therefore read as creating an additional requirement that is not reflected in the Article 514C and will be impossible to implement.
Regarding paragraph 33, we recommend preserving the original language of Article 51 of the first additional protocol that this paragraph draws from in order to ensure clarity regarding the legal obligation and to avoid further confusion on the implications of the deviations between this paragraph and Article 51 language.
Therefore, we suggest, I quote, the civilian population as such, as well as individual civilians shall not be the object of attack by laws.
Civilian objects shall not be the object of attacks by laws, end of quote.
Regarding paragraph 34, similarly to the request by other delegations, we also have serious concerns with this paragraph and we ask it to be deleted.
As in our view, the text should maintain its focus on the use of laws and not other actions regarding it, like manufacturing, stockpiling, et cetera As they stray from the CCW objective and expertise as an IHL focused treaty.
Regarding paragraph 39, consistent with our comments on paragraph five and 36, we will also suggest replacing reference to international law applicable law applicable IHL.
I will repeat this, sorry.
Consistent with our comments on paragraph 35 and 36, we will suggest replacing reference to international law with applicable IHL.
This will preserve consistency throughout the text and avoid introducing a broader legal standard in the context of laws.
Accordingly, the paragraph will read, quote, States must as appropriate across the life cycle of laws and in accordance with their obligations under applicable IHL, et cetera We will also support the US suggestion to remove must from the Chabot and to add new in paragraph A.
Regarding paragraph 40 B, we suggest deleting the word in before control of loss.
The term is unclear in this context as it breaks up the acceptable term of chain of command and control.
Furthermore, it may unnecessarily reintroduce the broader debate concerning the issue of human control.
The purpose of this provision appears more practical.
Ensure that the relevant personnel receive appropriate guidance, training, and instructions.
We therefore believe it will be clearer to refer simply to those in the responsible chain of command and control without introducing an additional concept of being in control of the system.
It will read, quote, ensuring appropriate guidance, training, and instructions for those in the responsible chain of command and control of laws.
Regarding paragraph 40 C, we continue to have serious concerns with the reference to explainable and traceable.
It remains unclear what precise standards of these terms established or how they relate to compliance with IHL.
We are also concerned with requiring the effects of loss to be predictable, which may suggest a degree of certainty that IHL does not require.
Therefore, we will suggest adding the word sufficiently or adequately before predictable in the first part of the sentence and deleting the second part of the sentence.
Therefore, it will read, quote, ensuring that laws in their identification, selection, and engagement functions operate in a sufficiently predictable and reliable manner.
Regarding paragraph 46, chair, we see some overlap with the preceding provisions on responsibility and accountability and will therefore support further streamlining.
If this paragraph is retained, we believe it should be more closely anchored to the subject of this group.
The phrase violations of international law, including IHL relating to laws is very broad and relating to does not purely identify the relevant legal obligations.
We therefore suggest referring specifically to AHL obligations applicable to laws, so it will read status must take all appropriate steps to prevent and suppress violations of IHL obligations applicable to laws by persons on the territory under their jurisdiction.
Finally, Chair, regarding paragraph 48, we believe that this paragraph may intentionally, create legal and practical ambiguity which its new placement may further complicate.
As we have noted previously, not all elements contained in the text have the same normative status.
Some reflect HL obligations while others function as practical tools or measures to facilitate the implementation or may reflect recommended practices.
As a general clause suggesting that all these elements should be taken into account in design, development, and use of laws, risk blurring these important distinctions.
In addition, the relevance of the different elements may vary between stages.
Some relate primarily to development while others concern the use of laws.
For example, it's not clear why provisions concerning legal revisions should also be taken into account at the use stage.
Over, the reference to other possible measures, it's open ended and undefined, which further increases the uncertainty as to what states are expected to take on account and what normative weight should be attributed to such measures.
For these reasons, Chair, we suggest deleting paragraph 48.
This is all for this morning.
Thank you so much, Chair.
I thank the distinguished representative of Israel for commenting on the text and making proposals.
Next on my list is distinguished representative of the Russian Federation to be followed by Belgium.
But let me first pass the floor to the distinguished representative of the Russian Federation.
Welcome.
Thank you, Chairman.
In light of the fact that other delegations have presented their considerations on paragraphs which had not been amended by you, in addition, we would also like to make a number of comments on those paragraphs.
Regarding paragraph 31, We still have the same concerns which we voiced regarding paragraph 33 in the context of the use of the terms it is prohibited.
In this connection, we would suggest removing the start of paragraph 31, particularly the sentence it is prohibited in all circumstances to use.
Instead of that, we would add at the end of this paragraph after the sentence or the part of the sentence in compliance with AHL must not be used.
On paragraph 32, along the same lines, we would suggest removing it is prohibited to use and instead, after the word laws, we would add must not be used.
Furthermore, in paragraph 32, we would like to suggest, as we did during the previous session, adding the wording of 33 A of the 2019 report, which was agreed and approved by consensus within this group of governmental experts and which relates to the stages of design, development, testing, and use of lethal autonomous weapons systems.
It reads as follows.
During the design, development, testing, and deployment of laws, the risks in of civilian casualties, as well as precautions to help minimize the risk of incidental loss of life, injuries to civilians and damage to civilian objects must be considered.
Other types of risks should be considered as appropriate, including but not limited to the risk of unintended engagements, risk of loss of control of the system, risk of proliferation, and risk of acquisition by terrorist groups.
Mia We consider this wording to be extremely important in the context of this prohibitive paragraph, as it is in the first sentence of Para 32, it explains the necessary requirements and possible restrictions which would be foreseen at the stages of design, testing, deployment of these weapon systems.
Furthermore, in paragraph 30 of the 2019 report, there is a direct link to international humanitarian law regarding the risks which are described in this paragraph.
Therefore, we consider it to be crucial to include this language that was agreed to by consensus in this document so as to not lose those elements because they are important in the context of delivering on the mandate of this group.
On paragraph 34, like a number of other delegations, we consider that this paragraph is superfluous and suggest cutting it in its entirety.
Because first of all, it contradicts para 39 where we speak of the legal reviews.
Moreover, this paragraph 34 creates the possibility For an interpretation of the situation when weapon systems would be created or designed and which could be used not in compliance with or in breach of international humanitarian law and which would not be adequately assessed or evaluated during the time of the legal review.
This paragraph foresees a situation when regardless of the need for a legal review, states would allow the possibility of developing and producing weapon systems which are prohibited under international humanitarian law and which should be assist as being not permissible for production or deployment as weapons at the stage of the conducting of the legal review.
This is just not possible because there are clear legal obligations to carry out such legal review so as to not allow precisely the development and production of such weapon systems that would breach international humanitarian law.
So in the context of paragraph 39, paragraph 34 appears to be nonsense.
And we consider it to be absolutely necessary to remove paragraph 34 as it contradicts paragraph 39.
Moreover, Chairman, We would like to suggest for paragraph 44, after the part of the sentence, individual accountability for, we'd like to suggest adding employment of because we consider that individual accountability applies also to the sphere of the use of the means or methods of warfare.
We consider that such an amendment would bring this paragraph into accordance with existing norms and principles of international humanitarian law and would make it possible for this language in this paragraph to be consistent in the context of existing obligations.
On paragraph 45, we consider it is necessary to correct the chapot In order to ensure that it relates to strictly possible measures that are brought up in the sub paragraphs of P 45 as measures that are voluntary, they're applicable on voluntary basis.
Given that we are speaking about measures that states could undertake at the national level, we do not believe it is necessary to prescribe in a mandatory way the necessity or the advisability of adopting such measures at the national level in an obligatory fashion, the set of measures could be different depending on the country specificities and the national practices that exist and imposing an adoption of all these measures would not be advisable in our view.
Therefore, Instead of the word should, we would suggest adding the sentence or the word may, may in on a voluntary basis provide.
In our view, this wording would allow for the possibility of the adoption of relevant national measures without predetermining or prescribing in a mandatory fashion what specific measures states could take on armaments at the national level.
On subparagraph B, we have a small or minor correction suggestion.
We continue to suggest removing the words rules of engagement that are in brackets because in the national practice of the Russian Federation, there are no rules as of such like this.
We have similar measures that are somewhat different and they are in other documents and other concepts.
Furthermore, we believe it would be justified for the word doctrine to be used in the plural doctrines.
As the other words in those brackets are policies and procedures are both in plural.
Thank you, Chairman.
I thank the distinguished representative of the Russian Federation for additional comments and reacting to proposals made by other delegations.
Next on my list is Belgium to be followed by Switzerland and Italy, but let me first pass the floor to distinguished representative of Belgium.
You have the floor.
Thank you very much, Mr.
Chair.
Thank you for giving me the floor.
At the outset, I would like to say that we are quite happy with the text you have proposed with the changes.
On paragraph 39 B, we would like to see it retained as it is for reasons that were very well explained by other delegations, Pakistan, ICRC, Brazil, New Zealand, among others.
We cannot consider the change proposed by New Zealand on the issue of substantial impact.
On Paragraph 40 A, we like very much the formulation that was proposed by ICRC to amend this Spa.
On 40 C, my delegation is very much attached to the notion of predictability, reliability, cability and traceability for reasons that were very well explained by previous interventions.
On paragraph 47, this is a very important para for delegation and we would like to keep the formulation must.
But this makes and we could accept a caveat at the end of the para in accordance with international law to make this must more logical.
On paragraph 48, we would like to see this para retained and from our point of view, it should be reinforced with a must or need to Um, but we heard concern about the dispara from other delegation, and we could live with a caveat to distinguish the different obligations that relate to the different phase of the lifecycle of the system.
So we could use a caveat on this 48.
We would like to see it retained.
Thank you very much.
I thank distinguished representative of Belgium for sharing the comments of his delegation and making proposals and reacting to the remarks of other delegations.
Next on my list is Switzerland to be followed by Italy, Sweden, and Austria.
But let me first pass the floor to distinct representative of Switzerland.
You have the floor.
Thank you, Mr.
Chair for giving me the floor.
Switzerland would like to briefly share some considerations regarding the following paragraphs that have been questioned by some delegations.
On Paris 30, 31, and 33, they follow widely accepted language stemming from existing treaties, including protocols of the CCW, notably amended Protocol two and Protocol three.
And apply them to laws.
We can thus support them as they stand, and we would caution to change their wording so as not to deviate from existing treaty language.
On paragraph 39, we support the proposed amendments and de alignment with the text of Article 36 of Additional Protocol one.
Together with the Chabot, we are very close to guiding principle E, which is now adapted to specifically address laws.
Switzerland cannot support proposals to delete paragraph B.
In our view, the further development of or modifications to existing weapon systems is not explicitly covered in paragraph A, which is why we consider the addition of paragraph B necessary for the same reasons highlighted by the delegation of Pakistan and others.
On paragraph 43, we support the proposed wording as it partially reflects guiding principle B and thus builds on language and principles that have already been agreed upon.
Paragraph 44, we appreciate that the wording remains close to the guiding principle D and consider that it reflects general IHL principles and principles regarding responsibility of the chain of human command and control.
We would, however, recommend staying closer to the wording of guiding principle D, in particular, that accountability must be ensured in accordance with applicable international law, not just IHL.
Paragraph 46, Switzerland agrees with this formulation as it is based on Article 4014, paragraph one, amended Protocol two, CCW and other IHL sources.
We do not see why this agreed language should raise concerns.
We welcome that as an element, the prevention and suppression of violations is extended to international law, at least until an instrument is adopted.
Paragraph 47, we appreciate this paragraph because we consider that effective domestic implementation mechanisms are an indispensable element for an effective accountability framework and for respecting and ensuring respect for IHL domestically.
The paragraph leaves open and leeway how exactly this should be done.
Lastly, my delegation welcomes paragraph 48, which takes up the different stages, notably design, development, and use.
I thank you, Mr.
Chair.
I would like to thank the distinguished representative of Switzerland for sharing the views of this delegation and reacting to proposals made by other delegations.
We have still 13 requests for the floor.
Let me now pass the floor to Distinguished representative of Italy to be followed by Sweden.
Thank you, Chair for giving me the floor and allowing me to comment on this paragraphs.
With regard to Par 39, we support the current formulation which emphasizes the need to ensure compliance with international humanitarian law through legal review mechanisms applied throughout the entire life cycle of lethal autonomous weapons systems, including in the event of any significant changes affecting their core functions.
With regard to paragraph 40 A, we remain flexible.
We support the new formulation while remaining open to possible simplification of the text.
At the same time, we consider it necessary to retain the reference to testing and evaluation, including simulations reflecting the operational environment prior to employment.
With regard to subparagraph C, we consider it important that lows perform all their primary functions in a predictable and reliable manner and that they functioning and in fact, be traceable and explainable.
We welcome the current rephrasing and remain flexible.
In our view, traceability and explainability are important safeguards for ensuring responsible use and accountability, and for preventing unintended engagements.
In addition, understanding how an autonomous system operates and the reaches an outcome as those responsible for its development and use to identify potential biases, assess their implications, and take appropriate measures to prevent and mitigate them.
Furthermore, explainability is necessary to mitigate unintended biases.
We recognize that these terms and the related legal considerations are understood differently and do not command consensus in the room.
Nevertheless, we remain flexible and ready to work towards a formulation acceptable for all delegations.
I thank you Chair.
Let me thank the distinguished representative of Italy for commenting on the text and reacting to remarks by others.
Next on my list is Sweden to be followed by Austria.
Let me now pass the floor to the distinguished representative of Sweden.
Thank you very much Chair.
Sweden appreciates and support many of the changes made in the additional suggestions.
We are also open to reformulations, provided that they do not significantly change the scope and the substance of these elements.
Generally, we prefer text that remains close to existing international law, reflects how it relates to laws and ensures that existing international law is not undermined.
As regards paragraph 34, we support the aim to prevent manufacture and proliferation of laws that are incapable of being used in line with international law.
However, similar to France and as we have previously stated, the current formulation of paragraph 34 causes us some concern.
We reiterate that this paragraph must not be understood as preventing the potential need to be able to deal with such systems in smaller quantities and for legitimate purposes.
For example, research for protective measures or accountability.
Legal review is a key issue for Sweden and therefore we consider that paragraph 39 A is an essential part of the text.
We welcome the changes made in paragraph 39 A of the chair's additional suggestions with a proposed insertion in the study, development, acquisition or adoption of laws, and we support the proposals to add the word new.
As others have mentioned, this reflects language that this group has already found consensus on both in the 2019 guiding principle E and in the 2023 report, para 23.
Notably, these formulations include both the word must and references to states obligations under international law.
We can accept the changes suggested for paragraph 40 and are open to reformulation of Paragraph 40 A and C.
But we would like to emphasize the need to keep subparagraph 40 as we consider these measures are important to ensure that laws be used in compliance with international law, in particular, IHL.
We support proposals to include laws in paragraph 41 and 42.
Finally, in regard to paragraph 46, we reiterate that we would like to see must be replaced with should since it goes beyond existing IHL.
We also support placing paragraph 48 at the end of the elements.
Thank you.
I thank the distinguished representative of Sweden for sharing the views and comments from her delegation.
Next on my list is Austria to be followed by Canada and Greece, but let me first pass the floor to representative of Austria.
You have the floor.
Mr.
Chair, we thank you for your tax proposal on this section, which is once again a sound proposal for a landing zone that we support very much.
The ICRC has made an important general point that we would like to re emphasize that we see our work here in the fulfillment of our mandate not only to repeat existing obligations of IHL, but also to clarify and further develop how IHL applies to autonomous weapons systems and their specific characteristics.
Years of hard work have been invested by this group in doing this.
Many comments and suggestions were made.
We will study part of them further and for the moment refer to some of them.
We heard comments on the prohibitions in paragraph 32, which not only has been there in the text for a long time, not put into question by any delegation, but also fully reflects existing IGL obligations enshrined in the basic principles of IGL.
It is hard for us to understand the concerns raised about this.
We attach great importance to this reference to remain as it is.
We likewise attach great importance to the prohibitions in para 33 and 34.
We would be prepared to look at technical language teaks in paragraph 33 if this helps.
We see Par 34 fully in line with paragraph 39 and appreciate the paragraph, following the logic of Article 36 that weapons systems that have been reviewed as not complying with IHL would not be acquired stockpiled or transferred.
And that we assess also has the necessary exceptions included to accommodate the concerns raised about small quantities for testing.
Umpire 39 itself, we stress the importance of fleshing out the elements of legal reviews in the text and ensuring continued regarding new developments or modification of existing laws as reflected in the chair's text.
The Chabot already includes a qualifier to refer to the legal obligations that apply for states.
There may be language tweaks to further reinforce this point and keep the essence as it is.
On paragraph 40 C, we attach great importance to the concepts of reliability, predictability, explainability, and traceability as safeguards reflected here, and we fully support keeping them.
We are open to look at language proposals to further accommodate the concerns we heard.
On paragraph 47, we think this paragraph leaves it up to states to decide on how to implement it.
We therefore believe the national prerogative is well taken into account here.
Finally, on paragraph 48, different suggestions have been made.
We see the need for a bottom line for a formulation that takes into account of the different categories of elements reflected here and particularly not to undermine the prohibition part of the text.
Thank you.
I would like to thank the distinguished representative of Austria for stating the positions of our delegation and making comments on the text.
Next is Canada to be followed by Greece and China.
But let me first pass the floor to the distinguished representative of Canada.
You have the floor.
Thank you, Chair.
My delegation would like to contribute to this discussion and has recommendations to make in five of the paragraphs that we're looking at.
Firstly, in Paragraph 34, Canada supports the intervention delivered by the Republic of Korea and supported by others for the reasons that have been outlined already.
Confirm that we also insist on the terms may or should be used rather than shall.
Jumping ahead to paragraph 38 A, Canada supports the objective of this subparagraph.
We have concerns with the reference to ethical considerations.
International humanitarian law does not impose an obligation to consider ethics in the operation of weapon systems.
As drafted, the phrase introduces an undefined normative concept that may blur the distinction between legal obligations and broader policy considerations.
Canada would therefore prefer to delete the reference to ethical considerations, but to demonstrate flexibility, my delegation would be open to introducing this concept in paragraph 45 B as an applicable guidance or instructions.
Excuse me.
Turning to paragraph.
We consider that the current wording could be more legally precise.
As drafted, the obligations for states to ensure individual accountability could be understood as requiring states to guarantee a particular outcome.
In our view, existing international humanitarian law requires states to establish and maintain appropriate legal and institutional mechanisms to give effect to individual accountability, including, where appropriate, through the investigation and prosecution of alleged violations.
We therefore suggest the following formulation.
States must take measures to ensure individual accountability for the means or methods of warfare involving the use of laws in compliance with their obligations under IHL, including through the operation of such systems within a responsible chain of command and control.
In paragraph 45, Canada supports the objective of this paragraph.
We consider that the drafting could be more closely reflect existing treaty language.
In particular, we would prefer the following formulation, take the measures necessary rather than take all appropriate steps.
As this is more accurately reflecting the language used in the Geneva Convention and additional Protocol one regarding the prevention and suppression of violations of IHL.
We also suggest considering whether the reference to international law, including IHL, should instead refer specifically to rules of international law applicable in armed conflict, consistent with the focus of this text.
Paragraph 46, Canada supports the objective of this paragraph as well.
We consider that it would benefit from closer alignment with existing treaty language.
Rather than creating an obligation to ensure effective internal mechanisms, prefer language that reflects states existing obligations to take the measures necessary to prevent, investigate, and respond appropriately to alleged violations of IHL.
This would more closely align to the provisions of the Geneva Convention in additional Protocol one.
Therefore, we recommend replacing take all appropriate steps with take the measures necessary.
Thank you, Chair and apologies for the length of my intervention.
Thanks.
Let me thank the distinguished representative of Canada for commenting on the text and providing concrete proposals.
Next on my list is Greece to be followed by China and Chile.
But let me first pass the floor to distinguished representative of Greece who have the floor.
Thank you very much, Mr.
Chairman.
At this moment, I will concentrate only on paragraph 39, where we would like to propose adding the wording, including international humanitarian law after the term international law, particularly in its first sentence in the Spot and in subparagraph A.
In our view, these additions would enhance the paragraph's coherence since subparagraph B refers to the need to ensure compliance with IAHL while no explicit reference to IHL appears at any earlier point in this paragraph.
Thank you.
I Distinguished representative of Greece.
Next on my list is China to be followed by Chile and Norway.
Let me now pass the floor to distinguished representative of China.
You have the floor.
Mr.
Chair, as to the text under discussion this morning, China has two points to make.
First, China suggests in 40 C, deleting three terms predictable, explainable, and traceable.
We should respect the sovereign jurisdiction of various countries, as well as their different national circumstances and military context.
In order to avoid overly penetrating or interfering in their domestic affairs.
We need to adequately address, on the one hand, the binding and compulsory nature of international rules and on the other hand, the autonomy and proactiveness of national regulation.
In so doing, we can ensure that the binding requirements are neither too loose or neutral nor too strict and they are easy to implement.
Second, As to the paragraphs containing the formulation must in the four Geneva Convention and the additional protocols and taking this as an example, must are mostly used in relevant provisions on specific circumstances or requirements regarding certain scenarios, rather than in the limiting national responsibility.
Just because the word must is not used will not necessarily lower the authority and effectiveness of relevant international instrument.
Therefore, China supports the suggestions by relevant delegations, Russia, US, Turkey, Israel in paragraphs 39, 44, 46 47 to change must into should.
Thank you, Chair.
I thank the distinguished representative of China for commenting on the text and making proposals.
Next on my list is Chile to be followed by Norway and Mexico.
But let me first pass the floor to distinguished representative of Chile.
Thank you, Chairman.
Before going into the paragraphs 39 onwards in detail, I'd like to set out a point of principle which guides my delegation's position on this entire section.
For Chile, International Humanitarian law is the basis of our work.
It is and must be and must remain our guiding star.
This points not only towards preventing violations, it points mainly towards compliance and states don't only have the obligation to prevent and punish violations of IHL.
They have a duty to ensure that their behavior and conduct during an armed conflict is at all times one that is compliant with the rules of warfare.
Preventing violations is the consequence of that duty, not its substitution and this needs to be fully reflected in the use of laws and does not allow for any other interpretations of IHL, which through technological autonomy, seek to dilute existing obligations.
Therefore, my delegation is not able to support wording which would reduce to a minimum the operative content of these paragraphs.
Now moving on to the specifics and paragraph 39.
And Spas A and B.
We support this section.
I do want to dwell on it a moment though because it seems to be particularly important to us.
The obligations of a legal review is not the only legal obligation when it comes to the study development, acquisition or adoption of laws.
According to Article 36 of AP one, any modification that alters the functioning or characteristics of a weapon requires a new revision, and this is covered in subparagraph B, which for us is therefore a non negotiable part of this paragraph.
We also insist on keeping the as appropriate in the Spot and also the words in accordance with their obligations under international law, and also in subparagraph A, we believe that we have to have the voluntary exchange of relevant practices and that has to be voluntary because it does not need to include sensitive material, confidential material, et cetera We support paragraph 40, in particular, subparagraph C, which means that the use of laws has to be in a predictable, reliable manner and that their functioning and effects are predictable, explainable and traceable.
This is essential in order to ensure compliance with IHL, and we would not wish for it to be reduced to technical or very broad terms.
In paragraph 41, we support the adoption of measures to mitigate as much as possible unintended bias, including bias stemming from datasets and artificial intelligence models, which are used by the system.
We believe this is a due diligence requirement and not an absolute guarantee for eliminating all bias here.
In paragraph 42 along the same lines, we support the adoption of measures to mitigate automation bias, that is to say the risk that individuals may attribute excessive trust to the informational results generated by automated systems.
One point of clarification here is that the existence of this risk cannot be considered per se a measure for legal determination accountability in a specific case.
Regarding paragraphs 43 and 44, we unreservedly support that autonomy should not transfer legal responsibility to machines, nor should it delegate it from states to individuals.
We want to make that clear.
In paragraph 43, it says that individuals that remain at all times responsible should not be read as being an objective form of responsibility because it requires the concurrence of all elements of applicable law, including subjective elements and With regard to the chain of responsibility, command and control, we believe that is a useful safeguard, but it should not be confused with the regime for individual liability responsibility or discipline.
We support the three measures contained in 45 as a minimum floor, not as a ceiling.
That is to say they are essential for people operating the system and guidance and instructions and training of personnel to understand the guidance and instructions.
Also, we need to show that the guidance and instructions are consistent with applicable IHL.
For us, these are preventive measures which seek to reduce the risk of unintended attacks, but it does not replace nor modify the legal requirements that govern the determination of responsibility in a specific case.
On paragraph 46, we support the text with two clarifications.
First of all, that the paragraph should not be read as creating new forms of individual responsibility, nor should it be read as altering the procedural legal aspects of existing legal systems.
This is an obligation of due diligence under the current framework, not a new regime.
Secondly, as we said, the legal framework of each individual state is the vehicle through which to give effect to this, but it must adjust to what has been agreed to under international customary law, first and foremost and not be secondary to that.
We also do support the fact that there should be internal mechanisms for investigation, reporting and appropriate action in response to incidents.
We would make two clarifications, though.
May involve violations of IHL does not imply that any or all use rather necessarily involves violations of IIHL the determination needs to be done on a case by case basis in accordance with applicable norms.
Secondly, need to ensure that the mechanism is not understood as a purely reactive mechanism following a violation that has occurred, but as part as the environment for ensuring continuous safeguards as we referred to at the start of this intervention.
Paragraph 48, we support it as it is, including the words should be taken into account.
We believe that should go unaltered.
Thank you very much.
Let me thank the distinguished representative of Chile for stating the views of the delegation.
Next on my list is the distinguished representative of Norway.
You have the floor.
Thank you, Chair.
As a package deal, we could support the paragraphs 26 34 and 39 48, sorry, as you've proposed in your additional suggestions, but we understand, of course, that we haven't arrived at that stage just now.
On paragraph 33, chair, the change that you've presented on the screen now is something that we can support.
With regard to the proposals made a while ago by the Russian Federation, we took note that they prefer changing the formulation.
It prohibited to must not be and if we understood correctly, the reason that was given was that prohibited did not appear appropriate.
But as we went on at length about in March, prohibited and must be are, in essence, just alternative ways of expressing the same deontic proposition.
So in our view, it's important, nonetheless, that we retain the word prohibited here because this most clearly reflects the two tiered approach of prohibitions and regulations that an overwhelming number of states here, have expressed their desire for.
On paragraph 34, we support the current formulation, but could also support the French proposal for a reformulation.
Of course, better to avoid opening it up if we can.
And we would also add that paragraph 34, which is an extremely narrow prohibition since it only kicks in at the manufacturing level, doesn't contradict paragraph 39 on legal reviews, but rather sort of complements it or works in tandem with it.
I think Austria was onto that.
I mean, if a legal review concludes that a weapon system cannot used in line with international law then paragraph 34 underscores that this must not then be manufactured.
I think others have mentioned that as well, but the word manufacturing was carefully chosen back in the day precisely because of those concerns that some have raised today regarding the potential need to design and develop small quantities for testing of countermeasures and such.
Regarding paragraph 39, I in your revised proposal, that is, the chat and subparagraph A.
I mean, we're not in a position to accept any weakening of this language from must to should because for us, legal reviews are a legal obligation.
We know, of course, that not all states are party to AP one, Article 36.
This was a much debated topic in previous sessions and in formals and that is precisely why the chapo is caveat with the words as appropriate and the their obligations under international law.
So, I mean, India made a proposal a little while ago, if we understood it correctly to sort of reorder the paragraph to highlight in the first bit that states might have varying obligations under agile I mean, we'll have to see it on paper to be sure that it's something we can live with, but it sounded like something that could possibly be workable.
But I'm not sure.
I mean, we would definitely prefer just keeping the formulation that is there also because if we understood the Indian proposal right, it doesn't really amount to any substantive change.
Regarding paragraph 39 B, this is particularly salient and useful in light of the quickly evolving nature of laws.
The US made a proposal there that sounded interesting, but I'm not sure that we caught all details of it, so we would need to see that text on paper, I think, in order to be sure.
Paragraph 40, Yeah.
I mean, this is a very important one for us, and others have expanded on that at length.
These are important measures to ensure respect for IGL.
They may not be direct obligations in IGL, but our purpose here is not to rehash IGL obligations verbatim, but to propose practical measures that contribute to the implementation of IGL.
And we believe that paragraph 40 does precisely that.
I would just add also that paragraph 40 is expressly couched in non binding terms, which in our opinion should allay the concerns of those who are concerned that we not create obligations here that might be new compared to current law.
Regarding 40 A, we took note of the proposals made by the Russian Federation and then worked on by the ICRC.
I think we could work on those.
40 C and the concepts there, so predictability, explainability, and traceability.
I mean, the reason why we should retain those have been laid out again and again by others very well.
So I'll just refer to their points on this, Pakistan and Brazil, for instance.
And I would just add that we're not negotiating for the instrument here.
We're proposing some elements that can be built on and elaborated in future work.
And as we see, these concepts, which have been debated a fair bit across the years, provide enough guidance for them to be useful starting points for future work, even though we might not have an exhaustive common understanding of them yet, or I mean, exhaustive is perhaps not the right word because that's never possible to settle all interpretations of a concept or any concept.
As regards the proposal by France, I believe it was to add adequately or some other qualifier in this subparagraph.
I mean, we don't think this is strictly speaking necessary because a normal good faith interpretation of this language would not, in our opinion, entail that you need to have certainty.
But if it could allay some of the concerns that delegations have, we could consider that.
Paragraph 31 and 41 and 42 buys, I mean, both data bias and automation buys are very real risks in a We have an issue with your microphone.
Right.
My speaking time wasn't up.
I'll continue then.
41 and 42 and just bias, which is, as I said, a real risk in AI based systems and something we've also discussed very much over the years.
These two paragraphs are very important risk reducing measures that we attach great importance to.
We took of bias mitigation to reduction of risk to civilians.
We don't think that is the only purpose of bias mitigation and prevention, but we could be open to including this as an example.
Two more, and then I'll stop Paragraph 44, just so that we could not support changing must to should here.
I mean, the whole point of this paragraph is to avoid an accountability gap as machines take over tasks that were previously carried out by humans.
48, this is important to us, As well.
However, we understand the concerns that have been raised today by some, so we wonder if perhaps the following formulation could be a solution and it would then read the formulated elements and other possible measures are to be taken into consideration as applicable in the design, development, and use of laws.
The point of this is to underscore that this is our intention for how these elements are to be used and not an obligation that we intend to impose on states.
Thank you.
A question now.
Have you finished your intervention because there's no sound anymore, but every now and then there's something with your microphone.
I'm not quite sure where it cut off, but you had your proposal regarding 48.
Yes.
Okay.
That was the last.
Yeah.
Thank you.
Okay.
Then let me thank you for reflecting on the proposals made and trying to offer some text to solve issues.
I I still have one, two, three, four, five, six, seven request for the floor.
We will continue this discussion this afternoon and as I stated before, we will reconvene at 4:00, so you have a little bit longer lunch break also to facilitate informal consultations among delegations.
With regard to the comments so far, let me just make a few remarks.
I Yes, one of the issues we have been facing all the time is that we have different obligations under international law and international humanitarian law, in particular, among delegations, and we have always tried to reflect that by applicable, appropriate and words like that.
Where we have not been consistent, where we have not been entirely clear, we will have to do something about it, of course, because we have to get that right.
Um On the other hand, we cannot fiddle with existing obligations.
So there we have made certain choices with regard to the wording, and it will be difficult to change that because otherwise we will change obligations of state parties.
And let me also Again, refer to our mandate with regard to whether we are creating new obligations here.
By now, you all know it by heart.
So please reflect again on that.
Whether that is indeed the case or not.
No, it's not the case because we are not creating a legally binding instrument, we are creating some set of elements.
Of an instrument and then, et cetera.
Regarding the discussion about what I call the abilities, we have had a long discussion on it for quite a number of delegations.
This is something very dear.
Others have some problems in accepting it, in understanding it.
There we clearly have to do some work.
The chair very much appreciates some of the suggestions made from the floor regarding rephrasing.
We will have a look at that.
Um, so clearly, there's some homework to be done.
On the other hand, we have to be recognizant of the limited amount of time we still have this week.
And I consider the paras we have been discussing today also substantial.
We have, of course, had a very clear focus on the characterization and the paras, what is it, 35 38 regarding human element, human intervention.
But of course, this is a set of elements, and there is more on the bone than just those two.
So having said that, I wish you a good lunch break, looking forward to seeing you back at 4:00 in this room.
I have, of course, an announcement made regarding side events.
That's organized by Stop Killer Robots on autonomous weapons systems in conversation with investors.
With the participation of leading investors and asset managers in the room, this side event will explore how the financial sector can help shape, constrain, and benefit from a clear and enforceable global rule book autonomous weapons, I would say.
This side event is open to all, will be held today 115-245 in Room Concordia one in Building A of the Palet.
Room Concordia one, building A of the Plat.
See you at four.
M&E
Meetings & Events
5th Meeting - 2nd Session Group of Governmental Experts on Lethal Autonomous Weapons Systems 2026
Group of Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous Weapons System Geneva, 2-6 March and 31 August - 4 September 2026
Description
Provisional agenda
1. Opening of the meeting.
2. Adoption of the agenda.
3. Confirmation of the rules of procedure.
4. Organization of the work of the Group of Governmental Experts.
5. Further consider and formulate, by consensus, a set of elements of an instrument,
without prejudging its nature, and other possible measures to address emerging
technologies in the area of lethal autonomous weapon systems, taking into account the
example of existing Protocols within the Convention, proposals presented by High
Contracting Parties and other options related to the normative and operational
framework on emerging technologies in the area of lethal autonomous weapon
systems, building upon the recommendations and conclusions of the Group, and
bringing in expertise on legal, military, and technological aspects.
6. Other matters.
7. Adoption of the report.
8. Closing of the meeting.
Full transcript en transcript
Machine-generated · not human-reviewed · verify against the official record before citing or relying on this transcript
Session Summary Auto generated from session transcript
Synthesis hasn't been generated for this session yet.
The summarize pipeline runs after the English transcript is available.
Machine-generated · not human-reviewed · verify against the official record before citing or relying on this summary