Symposium on the CCW GGE LAWS: Obstacles Avoided? Rolling Towards Consensus for the 2026 GGE LAWS Report

[Jeroen van den Boogaard is a legal counsel for the Dutch Ministry of Foreign Affairs and a lecturer in international humanitarian law at the University of Amsterdam. He writes this post in his personal capacity. This is the seventh and final piece in a series of analyses on aspects of the recent adoption of the 2024-2026 CCW GGE LAWS report.]

In a piece published 4 March 2024, just before the first meeting of the current mandate of the Group of Governmental Experts on Lethal Autonomous Weapons Systems (GGE LAWS), I identified three potential obstacles for the regulation in the 2024-2026 mandate of the GGE LAWS. As a member of the team assisting the GGE LAWS Chair, Ambassador Robert in den Bosch, I had the privilege to witness first-hand how the final Report of the 2024-2025-2026 sessions of the Group of Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous Weapons Systems came to its conclusion in the early hours of Saturday 5 September 2026.

In this post, I first reflect on the obstacles identified in 2024, the process of the GGE LAWS since then, and on the outcome in the Report that will be offered to the CCW Review Conference taking place in Geneva from 16-20 November 2026. In doing so. I will also reflect on the analyses of the different parts (formerly known as “the boxes”) of the 2026 Report in the previous posts in this week’s Symposium.

Obstacles Ahead: The Mandate is the Mandate is the Mandate.

A first obstacle identified in 2024 was that there were both States ready to negotiate and conclude a treaty prohibiting LAWS as soon as possible, and States that did not (at all). Therefore, one of the challenges during the mandate was to focus on, and only focus on, what the mandate provided: to “further consider and formulate, by consensus, a set of elements of an instrument, without prejudging its nature, and other possible measures.”

Since the GGE LAWS had a three-year mandate, and its Report was not due until 2026, the Group was able to focus on making substantive progress, and to overcome the obstacles in regulating autonomous weapons systems. This was the approach throughout: there was no general debate and procedural issues were handled swiftly.

The approach of the Chair to identify the “elements” in the mandate was to divide these in different themes, framed in 5 different boxes of a Rolling Text, implementing a two-tier approach of prohibitions and regulations. These themes were (I) a characterization of lethal autonomous weapons systems; (II) a list of International Humanitarian Law (IHL) rules and principles particularly relevant to autonomy in weapons systems; (III) prohibitions and regulations; (IV) other measures and (V) responsibility and accountability. Although these boxes had these titles in the first versions of the Rolling Text, the Chair decided to remove these, because discussions on the title of the boxes seemed rather pointless, and were hampering discussions on the actual substance of the boxes.

The Rolling Text was regularly updated throughout the mandate, but with small steps. The Chair reminded delegations time and again that the entire Rolling Text, that was eventually turned into the (draft) final Report near the end of the process, was a package deal, on which States needed to find consensus.

Characterization: Belts and Braces

A second obstacle, that of a definitional nature, was to establish what we had actually been talking about since the inception of the GGE LAWS. What is autonomy, and in what way is an autonomous weapon different from any other weapon? And what does it mean that we are talking about weapons “systems”? And what does the qualifier “lethal” in the mandate and the term LAWS really mean in the context of IHL?

The characterization (the term “definition” had become tainted much earlier in the GGE LAWS process) of LAWS proved to be a subject that still required much debate, perhaps more than anticipated at first by the Chair’s team. In my view, the sticky problem here was that there were two camps of States. Some States were comfortable with agreeing to a broad characterization of the systems under discussion, whereas others were not. The latter group of States seemed to be striving for a very narrow characterization, preferably excluding weapon systems they already had or were in the process of developing. It proved necessary to use increasingly more words to get it right, which in turn sparked new discussions. And it proved necessary to keep the debate on the adoption of an agreed characterization of LAWS open until the very end, so it could be part of the package deal on that late Friday night in September 2026.

The solution to many of the challenges faced in finding common ground on the characterization was to explain time and again what the terms used meant, and to provide separate paragraphs explaining some of the terms in the characterization on top of that. But that explanation is, admittedly, difficult to follow. Rain Liivoja’s eloquent analysis of para 25 of the Report, formerly known as Box I, is highly recommended therefore, because it provides a much-needed “explainer of the explanations” found in paras 25a (on identification) and b (on lethality).

IHL as It Is, Should Be or Could Be?

According to its preamble, work within the framework of the CCW unavoidably should serve “the need to continue the codification and progressive development of the rules of international law applicable in armed conflict.” It was therefore important to find a balance in the GGE LAWS between maintaining a progressive attitude, while at the same time, adhering to the mandate of identifying elements. In that process, safeguarding the existing rules proved challenging at times, and was further complicated by the fact that not all States are party to the same IHL treaties, and thus have differing legal obligations. The two-tier approach that steered the GGE LAWS in the last few years proved helpful, but also gave some rise to discussions. A mere restatement of existing IHL obligations in the elements would not contribute much, and changing words or even commas from established IHL rules into the GGE Report would only add confusion. Some States seemed to fear that agreeing on a “new standard of human judgement and control” in the GGE Report would lead to new international legal obligations. But there were also States that thought that repeating existing IHL prohibitions in the GGE LAWS Report using the words “it is prohibited” would be inappropriate. As Sally Longworth explains in her analysis, the rationale for this position is that “the GGE was not mandated to negotiate a treaty.”

Nonetheless, Box II on IHL received relatively few comments during the sessions of the GGE LAWS, and seemed agreeable as is to all delegations. The exception however proved to be the paragraph noting that “It is prohibited to use LAWS if their effects in attack cannot be anticipated and limited, as required by IHL in the circumstances of their use.” This seems striking because autonomous features of weapons systems would seem to make this prohibition particularly relevant for regulating autonomous weapons systems. Also, as Longworth notes, the text of this paragraph “clearly mirrors existing IHL obligations.” Nonetheless, instead of trying to find a formulation that would accommodate the last-minute objections (and proposed fixes for it) by some States, it was decided by the GGE LAWS Chair that it would be better to delete this paragraph than to end up with a paragraph that would dilute existing IHL. Also, the last-minute objections made it impossible to thoroughly discuss this issue on 4 September 2026.

MHC=C-AHC&J=C-AHJ&C=Control and Human Judgement

The third, and main obstacle mentioned in my March 2024 blogpost is the human element. How does one square human involvement in weapons systems that have autonomous features? As I argued, the relevant elements here are human judgement to ensure IHL compliance; control to ensure predictability, and an assessment of the context, because that delineates the extent to which these are necessary.

In the GGE LAWS, delegations agreed with the Chair that identifying the human element and finding consensus on its content, would be the most challenging part of the discussion. As a result, the majority of hours of discussion in the GGE LAWS was spent on this issue. The Rolling Text abandoned a previous formula, that of “meaningful human control” (MHC). Although many states and civil society organizations saw MHC as the solution to the regulation of autonomous weapons systems, and invested heavily in this term, the Chair saw the formula as tainted, and therefore a potential obstacle to finding consensus on the human element.

The first meeting under the mandate in March 2024 was preceded with a letter from the Chair with a number of guiding questions, and contained one question on this issue. Based on the responses received, and the discussions during the first week, the first Rolling Text (of 24 July 2024) contained two paragraphs with regard to control and human judgement that were previously agreed by the GGE LAWS as “preliminary considerations”:

Control with regard to weapon systems is needed to uphold compliance with international law, in particular IHL, including the principles and rules of distinction, precautions and proportionality.

Human judgement is essential in order to ensure that the potential use of weapon systems is in compliance with international law, and in particular IHL.

Further in the text, the phrase “appropriate control and human judgement” also featured.

The Rolling Text of 8 November 2024 first coined the phrase “context-appropriate human control and judgement”, which was subsequently, as Anna Andersson explains, changed to context-appropriate human judgement and control (C-AHJ&C) in the 12 May 2025 version of the Rolling Text. This inversion aimed to introduce some constructive ambiguity in order to accommodate those States fearing that “human control” as s new legal standard for the use of autonomous weapons systems. But in essence, the Chair’s team viewed these terms as interchangeably.

Although many States continued to prefer the term of MHC, others objected because they viewed it as yet another attempt to define a new legal standard applicable to the use of autonomous weapons systems. Some Sates however embraced this formula of C-AHJ&C as an appropriate framing of the human element. In a Working paper of August 2025, the so-called “Two-tier Group” provided a description of how C-AHJ&C could be understood (full disclosure: I was involved in the drafting of this paper).

Eventually, the GGE LAWS was able to find consensus on paragraph 35 of its final report, that contains a description of how “control and human judgement” works to uphold compliance with IHL. New in the paragraph is the observation that “control” can be exercised by humans “directly or indirectly, including through measures taken before and/or during the use of LAWS”. This proved an important breakthrough in the discussions. The reference to the context is contained in a separate paragraph 36 as “a range of factors must be considered, as feasible, including the operational context and the characteristics and capabilities of the weapon system as a whole.”

Box IV-V: Measures, Responsibility and Accountability

The subjects of Box IV and V were not mentioned in my 2024 blogpost as potential obstacles. In fact, there was already quite some language in previous agreed reports of the GGE LAWS on these issues. However, since the majority of time during the 2024-2026 mandates was spent on finding convergence on the LAWS’ characteristics and formulation of the human element, these subjects received less attention during the meetings of the GGE LAWS than they could have, and perhaps should have.

As Netta Goussac notes, some measures that never made it to the finish line of the final report would have had real added value. Even a pretty basic measure, such as limiting the number of engagements a LAWS would be allowed to undertake autonomously, would have been a truly practical measure for militaries, and was never to be understood as creating a new legal standard of IHL. Also, capturing the discussion on the “substantial” or “material” modification of LAWS into the Rolling Text could have been quite useful. And finally, it may be regretted that the so-called “ilities” were eventually deleted from the Rolling Text, since these notions (of predictability, explainability, traceability and understandability) could be of great practical assistance to users of LAWS.

With regard to responsibility and accountability, I agree with Jessica Dorsey that as a result of the limited time dedicated to Box V, the GGE Report “retains the idea that human responsibility cannot be transferred to machines, but provides considerably less direction on how that can be translated into meaningful accountability mechanisms in the context of LAWS.” Of course, whether more time would also have resulted in stronger language, is something that may be looked at in the future.

Key Elements of the Process Leading to Consensus

Looking back, in my view, we should treat the consensus on the 2026 GGE Report as a success. There are many factors that contributed to that success, such as the use of a “Rolling Text” format. There are also many people who contributed, not least the Chair himself, and the indispensable Friends of the Chair and the UNODA / ISU staff members. They were truly force multipliers. Furthermore, all participating delegates that gave so much effort and dedication (and hours!) in the meetings in Geneva, the informals, the intersessionals and the preparation for all of these. But it was a much broader team of people that made the Report possible (and it would be impossible to list all of them here). I am convinced that the thorough and expert analyses in the other post of this Opinio Juris symposium provide a solid foundation for readers to join me and all others on the journey. The next stop is the November 2026 CCW Review Conference, where the 2026 GGE LAWS Report will be presented.