Policy Brief No. 241 — July 2026
Time for Canada to Take a Stand on Lethal Autonomous Weapons Systems
Craig Martin
Key Points
→ Canada has maintained an understandable posture of ambiguity on lethal autonomous weapons systems (LAWS), particularly within the Convention on Certain Conventional Weapons (CCW) process.
→ Canada should now take a stand within the CCW and support the negotiations based on a “rolling text” — a proposed multilateral framework for governing LAWS that is the most developed yet produced.
→ The substance of the rolling text is closely aligned with Canada’s stated positions on ensuring meaningful human control, as well as insisting that all aspects of the development and deployment of LAWS must be subject to, and comply with, international humanitarian law (IHL). Canada’s substantive contributions to the Group of Governmental Experts (GGE) within the CCW — including working papers on algorithmic bias and IHL-based regulatory architecture — and its legacy in humanitarian treaty-making position it well to lead on shaping the proposed prohibitions and regulations.
Introduction: The Issues
Since the early 2010s, there have been debates over whether autonomous weapons systems (AWS) can fully comply with IHL, whether existing IHL treaty regimes can adequately govern the operation of AWS, and what form any new constraints on AWS should take. The primary forum for interstate discussion of these issues has been within the institutional framework of the CCW, and since 2016, by the GGE within the CCW.
For many years that consensus-based process showed little hope of making any meaningful progress. Canada’s own approach to the issues has been characterized by studied ambiguity. On the one hand, it has articulated support for a ban on fully AWS and participated substantively in the GGE process. On the other hand, it has avoided any concrete support for the negotiation of new binding international obligations or even enacted any meaningful domestic constraints on military artificial intelligence (AI) adoption and AWS development at home.
There are compelling reasons for Canada to now take a stand and even exercise leadership in resolving the issues regarding the governance of AWS. First, matters are coming to a head within the GGE. It was given a mandate by the CCW in 2023 to formulate a “set of elements of an instrument” to address the issues relating
About the Author
Craig Martin is a professor of law, and co-director of the International and Comparative Law Center, at Washburn University School of Law. He is also a graduate of, and frequent visiting lecturer at, Osaka University Graduate School of Law and Politics. He can be reached at craigxmartin@gmail.com.
to LAWS (Meeting of the High Contracting Parties to the CCW 2023). That mandate is set to expire at the CCW’s review conference in November of this year. Many see this as the last chance for the CCW to develop any significant constraints on LAWS.
Second, there have been meaningful developments within the GGE that suggest that there is some chance of success in achieving a resolution. In September 2025, Brazil led a group of 42 states in supporting negotiations on a “rolling text” as the basis for a new instrument, and in March 2026, the number of states supporting this position grew to over 70 states, which is more than half of the states parties to the CCW (Varella 2026). There is thus an increasing chance that the rolling text, which represents the most developed multilateral framework for governing LAWS yet produced, could form the basis of a new legally binding agreement. What is more, the positions advanced in the rolling text are substantively aligned with the positions that Canada has articulated as being crucial with respect to the governance of LAWS and are designed to better ensure that the development and deployment of LAWS will comply with IHL.
Finally, while there are few other existing fora for quickly developing a new governance structure for AWS or adjusting the existing IHL regime to better govern AWS, the pace of technological advance is rapidly overtaking the process. AI targeting and AWS with very limited meaningful human supervision have arguably already been deployed to varying degrees in hostilities in Gaza (Abraham 2024; United Nations Human Rigths Office of the High Commissioner 2024), Libya (United Nations Security Council 2021), Ukraine (Wells 2026; Bondar 2025), and even in the conflict against Iran this year (Nature Editorial Board 2026). The International Committee of the Red Cross (ICRC) and a growing number of legal academics and policy makers have advanced strong claims that AWS with little or no human supervision or control will not be capable of complying in all circumstances with the core IHL principles of distinction, proportionality and precautions in attack, and that new binding AWS-specific rules are thus necessary (Martin 2025, 261–64, 284–98).
This policy brief will briefly expand on these points and lay out a set of recommendations for why and how Canada should support and engage with the negotiations on the rolling text with a view to establishing a binding agreement on LAWS.
Canada’s Ambiguous Position on LAWS
The 2019 Ban Mandate and Domestic Divergence
Prime Minister Justin Trudeau issued a mandate letter in 2019 directing the minister of foreign affairs to “advance international efforts to ban the development and use of fully autonomous weapons systems” (Prime Minister of Canada 2019). This was viewed as a significant shift in position. Canada had, in the 2017 defence policy Strong, Secure, Engaged, committed the Canadian Armed Forces to maintaining “appropriate human involvement in the use of military capabilities exerting lethal force” (Department of National Defence [DND] 2017). At the same time, however, it had also aligned itself with the US position in the GGE that any ban on LAWS would be premature (Lee 2021). Even after the mandate letter, Canada’s position created ambiguity. Civil society critics noted that the “ban” language was not reflected in any Canadian statements in the GGE, and even military professionals suggested that Canada appeared to be hedging its position on AWS constraints (Marijan 2022; Lee 2021).
Domestic policy developments also appeared to widen the gap between the intentions reflected in the mandate letter and actual practice. The Department of National Defence and Canadian Armed Forces Artificial Intelligence Strategy (2024) committed the Canadian Armed Forces to becoming “AI-enabled by 2030” and identified robotics and autonomous systems as key capabilities, albeit while emphasizing that applications involving lethal force must retain humans in the loop (DND 2024b, 6–9). The defence policy update in 2024, Our North, Strong and Free, noted that AWS and AI are “redefining conflict” and allocated CDN$73 billion over 20 years in defence, including in AI, drones and autonomous technology, with no reference to developing governance structures to regulate AWS (DND 2024a). Bill 27, which included the Artificial Intelligence and Data Act, even excluded any application of the act to any AI systems developed or used for national security or national defence, though it died on the order paper in January 2025. There have been no other moves to enact domestic constraints on the development and deployment
of AWS or AI-targeting systems. Canada’s National Artificial Intelligence Strategy: AI for All, released on June 8, 2026 — while referring to collaboration with allies on “defence AI” and developing “dual-use applications” to align with national security commitments — does not address AWS or AI targeting systems (Innovation, Science and Economic Development Canada 2026).
In short, Canada appears to be trying to maintain a fine balance, gesturing toward endorsing a ban on fully AWS, and support for ensuring meaningful human involvement in any LAWS, while keeping its options open for developing and deploying nonfully AWS and military AI systems.
Engaged but Noncommittal in the GGE
Canada’s ambiguous hedging posture is also reflected in its activity within the GGE. Canada co-sponsored a working paper (GGE in the Area of Lethal Autonomous Weapons Systems 2024a) on the potential for algorithmic bias in LAWS with a group of states that are inclined to support strict governance of LAWS. At the same time, it sponsored two separate versions of a working paper that provided draft articles on IHL-based prohibitions and regulations on LAWS (GGE in the Area of Lethal Autonomous Weapons Systems 2024b, 2025b) with the United States, United Kingdom, Australia and several other states that are the most resistant to the development of binding legal obligations. While both of these working papers inform the substance of the rolling text and leave Canada well positioned to play a greater role (as discussed further below), Canada’s alignment with groups with quite different perspectives and positions may again reflect its hedging approach.
Similarly, Canada has voted in favour of the three UN General Assembly resolutions specifically addressing governance of LAWS,1 and in its November 2025 statement to the High Contracting Parties at the GGE, it affirmed that “compliance with IHL is not optional,” and again called for “ensuring meaningful human control” where it is necessary to uphold the core principles of distinction, proportionality and precautions in attack (Meeting of the High Contracting Parties to
1 United Nations General Assembly, Resolutions on Lethal Autonomous Weapons Systems, GA Res 78/241, UN Doc A/RES/78/241 (22 December 2023); GA Res 79/62, UN Doc A/RES/79/62 (2 December 2024); GA Res 80/57, UN Doc A/RES/80/57 (5 December 2025).
the CCW 2025). Yet Canada did not join the group of 42 states — which included such like-minded states and allies as Denmark, France, Germany, New Zealand and Sweden — that declared its readiness to negotiate an instrument on the basis of the rolling text (Stop Killer Robots 2025).2 Nor was Canada among the states that joined this group in early March 2026, expanding its number to over 70 countries.3 Automated Decision Research, a civil society organization that monitors the GGE process, classified Canada as being among the group of 54 states that are “undecided” on support for a legally binding agreement at the end of 2025, in contrast to 129 states that now support a legally binding obligation and 12 that openly oppose it.4 Yet, in May 2025, Canada co-signed an Australian-led statement cautioning that any effort to initiate a LAWS governance process outside the GGE would be counterproductive (Permanent Mission of Australia to the United Nations 2025).
In short, therefore, Canada espouses support for constraints on fully autonomous LAWS, and for ensuring meaningful human control and IHL compliance; it appears engaged in the GGE process and actively contributes to deliberations and working papers within it; it advises against any efforts to develop LAWS governance structures outside of the GGE — and yet it remains ambivalent about the development of any legally binding obligations within the GGE and has declined to support negotiations on the basis of the rolling text. The time has come for Canada to take a stand in support of the rolling text and a global agreement governing LAWS.
Explaining the Reluctance
The Government of Canada has not publicly articulated an explanation for its studied ambiguity in the broader deliberations or its reluctance to endorse the current rolling text. Some inferences can be made, however, from the government’s statements and submissions within the GGE, as well as from the dynamics characterizing its relations with Canada’s allies and even among its
2 The joint statement by Brazil, representing the 42 states, is available here: https://reachingcriticalwill.org/images/documents/Disarmament-fora/ ccw/2025/gge/statements/5Sept_Group.pdf.
3 Confirmed in a personal communication from Laura Varella of Reaching Critical Will on March 30, 2026 (Varella 2026). The records of the CCW/GGE do not yet reflect the specific states within the group.
4 See https://automatedresearch.org/news/state_position/canada/.
own ministries. One significant statement was provided by Canada in its May 2024 submission to the UN Secretary-General, advancing the socalled legal-sufficiency argument, in which it maintained that it was “unclear at this stage what gaps in the current international framework a new instrument would seek to fill.”5 One might be tempted to view Canada’s reluctance as being grounded in legal principle. This position, however, is at odds with Canada’s role in co-sponsoring a working paper with IHL-based prohibitions and regulations of LAWS, which suggests that Canada’s support for the legal-sufficiency argument does not reflect a considered legal position, but is merely a placeholder. It illustrates the ambiguity more than sheds light on it.
A more likely explanation for the ambiguity lies with alliance posture. That is, Canada’s studied ambiguity is more likely for the purpose of maintaining some alignment with the United States both bilaterally and as part of the Five Eyes6 and the groups forming the working paper WP.10/WP.7 coalition within the GGE. The United States and Australia, in particular, have taken strong stands against the rolling text being developed into a binding treaty, and the Government of Canada has in various recent circumstances exhibited extreme caution in taking any positions that might provoke or estrange the current American government.
An additional factor is that, as Project Ploughshares has observed, “it appears that Global Affairs Canada and DND have different perspectives on this subject” (Marijan 2022). The potential tension between these two groups in trying to pursue these differing perspectives may explain some of this ambiguity and particularly why the language of the 2019 mandate letter, which called for an advance of international efforts to ban fully AWS, has never been repeated. As noted above, the government’s recent emphasis on defence spending and development has included programs for AI and AWS, all of which may have strengthened the hand of the DND in this internal debate.
5 United Nations, Secretary-General, Measures to Uphold the Authority of the 1925 Geneva Protocol, UNGAOR, 79th Sess, UN Doc A/79/88 (2024), Annex (Canada) at 32, online: <https://docs-library.unoda.org/ General_Assembly_First_Committee_-Seventy-Ninth_session_ (2024)/A-79-88-LAWS.pdf>.
6 The Five Eyes is an intelligence alliance consisting of Australia, Canada, New Zealand, the United Kingdom and the United States.
The GGE, IHL and the Rolling Text
The 2023 Mandate
As mentioned earlier, deliberations in the CCW were making little progress until 2024. But at the Conference of the High Contracting Parties in 2023, the parties established a mandate for the GGE, directing it 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 weapon systems” (Meeting of the High Contracting Parties to the CCW 2023, paras. 20 and 26). This was considered a significant departure from the open-ended deliberations to date, though it left open whether the “instrument” would be a legally binding agreement or some lesser political declaration. The GGE was given until the review conference in November 2026 to complete this mandate.
In August 2024, the chair of the GGE introduced the rolling text to form the basis of the negotiations of that instrument, marking the first time that the GGE would be working on a consolidated draft document. It has been revised six times thus far, most recently in December 2025 (GGE in the Area of Lethal Autonomous Weapons Systems 2025c). As mentioned above, over half the members to the GGE now support negotiating an agreement based on the rolling text.
The IHL Foundation
The rolling text, and indeed the broader deliberations of the GGE, has been grounded in the premise that IHL applies to LAWS and that LAWS must comply with IHL. Where there has been some disagreement is over the question of whether LAWS can always comply with IHL and whether IHL needs to be updated to better govern LAWS. Many states — and indeed legal scholars and organizations within civil society, including the ICRC — have argued that LAWS will not be able to fully and reliably comply with the core principles of distinction, proportionality, and precautions in attack without sufficient meaningful human involvement and control. This is particularly the case in operational environments
in which there is significant civilian presence, or where combatants or fighters are difficult to distinguish from civilians — such as hostilities in populated urban areas or in non-international armed conflicts. Thus, there are calls for limiting the geographic and situational scope of permitted operations of LAWS (Martin 2025; Winter 2022).
The reasons for these calls are multifaceted and complex, grounded in both technical aspects of AWS capability and the deeply contextual human judgment required to implement the core principles of IHL. For instance, while pattern recognition is a particular strength of AI platforms, machine learning (ML) models are prone to significant error when the population or environment they operate in diverges from their training data, and these errors can proliferate in non-linear feedback loops and develop into systemic biases (Narayanan and Kapoor 2024; Ferrara 2024; Manson 2026). In the targeting context, this means that an AWS trained on one population but operating in a different culture and environment may systematically misidentify civilians and fail to distinguish fighters from the civilian population (Winter 2020; Martin 2025).
Similarly, AI has demonstrated particular weaknesses in predictive decision making, with studies showing that many predictive models perform little better than chance when applied to complex, contingent situations (Narayanan and Kapoor 2024). Yet this is precisely the capacity that the principle of proportionality demands: the assessment of expected civilian harm weighed against anticipated direct and concrete military advantage. Both sides of the equation require forward-looking judgments about abstract values being impacted by evolving and uncertain conditions (Martin 2025).
Moreover, AWS are vulnerable to jamming, spoofing, and cyberattacking that may sever the continuous intelligence feeds essential to sound targeting decisions, or even corrupt or co-opt the decision-making process itself (Scharre 2019). This vulnerability has been demonstrated in conflicts such as Ukraine, has proven difficult to defend against (Khlaaf, Myers West and Whittaker 2024; Kwik 2024) and has the potential to interfere with the operationalization of all core principles of IHL. In addition, the “black-box” problem — the inability of even the designers of ML systems to explain or predict how or why targeting decisions are made —
raises fundamental challenges for both compliance with the core principles of IHL and accountability in the event that potential violations occur (ICRC 2020; Boulanin, Bruun and Goussac 2021). It is precisely because of these and other perceived problems that many states argue that there must be “meaningful human control,” leading to debates over the extent to which, and in what circumstances, humans may be “in the loop,” “on the loop,” or even entirely “out of the loop” in some circumstances.
The Rolling Text
To ground the recommendations that follow, it is necessary to explain some of the details of the rolling text, which is organized into five thematic “boxes.” Box I provides a definition of LAWS that the proposed instrument would cover: “a functionally integrated combination of one or more weapons and technological components, that can identify, select, and engage a target, without intervention by a human operator in the execution of these tasks” (GGE in the Area of Lethal Autonomous Weapons Systems 2025c). It includes elaborations that partly reflect the long-standing debate within the GGE over the definition of LAWS. This includes disagreement on limiting the focus to “lethal” systems, and even what “lethal” entails, but also extends to whether AWS include those that “identify,” as well as “select and engage,” targets without human intervention.
Box II reiterates that IHL applies to all means and methods of warfare, which thus includes the development and deployment of LAWS, and that the choice of means and methods of warfare are limited by IHL. It further invokes the famous Martens Clause to emphasize that persons remain under the protection of the principles of humanity even in situations not covered by specific treaty rules.
Box III is the substantive core of the rolling text, which can be divided into two tiers. The first establishes certain prohibitions on specific forms of LAWS — namely, systems that are inherently indiscriminate, cause superfluous injury or unnecessary suffering, or whose effects in attack cannot be anticipated and limited as required by IHL, particularly in terms of the principle of distinction. Notably the prohibitions extend beyond use to cover manufacture, acquisition, stockpiling and transfer of non-compliant systems.
The second tier relates to the establishment of a regulatory framework to govern non-prohibited LAWS, which would include a requirement that
LAWS comply with, and be used in accordance with, the core principles of IHL. Paragraph 6 relates to the central issue of the degree of human involvement or control that should be required of LAWS. Canada has been one of many countries that have consistently advocated for requiring “meaningful human control,” while the United States and others have resisted any binding human control standard. The compromise formulation, reflected in paragraph 6, is context-appropriate human judgment and control or CAHJC (GGE in the Area of Lethal Autonomous Weapons Systems 2025c). Significantly, this issue was recently moved from the more conceptual Box II to the substantive Box III in order to anchor the human-control requirement in the structure of IHL obligations.
Paragraph 7 lists several specific measures to be taken to ensure CAHJC as part of this tier of regulation. These measures include ensuring that LAWS always be operated under a responsible chain of command and control, including specific human assessment of legal obligations regarding target identification, selection and engagement. Another measure is to ensure that the effects of any use of LAWS can be adequately anticipated and controlled by those responsible; this includes restricting the types of targets, the duration of operations, and the geographical scope and number of engagements that LAWS can undertake autonomously. Yet another especially crucial measure is one that ensures that mission parameters of LAWS cannot be significantly modified by the system itself without human intervention and requires that LAWS can be deactivated in a timely manner. A final measure requires a limitation on the operation of LAWS to within defined perimeters and restricts their use to the targeting of military objects to minimize the probability of harm to civilians.
Box IV provides for the governance of LAWS over their entire life cycle and elaborates on the kind of Article 36 reviews that are required for LAWS, with the maintenance of human control throughout the life cycle. Box V relates to accountability issues and emphasizes that individual human accountability must be ensured throughout the entire life cycle of development and deployment of LAWS. Among other things, this requires ensuring an understandable human-machine interface.
As will be discussed next, the recommendations here for Canada relate most directly to the proposals in Box III.
Recommendations for Canada
Canada’s apparently ambiguous and ambivalent posture to date on LAWS, both within the GGE and more generally, is understandable. There is a tension between, on the one hand, its principled position regarding adherence to IHL and its legacy of leadership in negotiations on issues of disarmament and weapons limitation (the AntiPersonnel Mine Ban Convention being the most obvious example), and, on the other hand, its need to maintain relations and interoperability with allies and to retain the flexibility required to address its own national security requirements in a rapidly changing strategic environment. Nonetheless, given the looming deadline for resolving the issues on LAWS within the GGE — and the absence of any realistic alternative for developing a global governance system for LAWS if the CCW process fails — the time is right for Canada not only to take a more committed position but to exercise leadership in the process. The following recommendations would entail one possible path toward doing so.
Endorse the Rolling Text
Canada should join the more than 70 countries that have endorsed the rolling text as a basis for negotiating an instrument on LAWS. This would require no change in Canada’s substantive position, being entirely consistent with Canada’s insistence on adherence to IHL and maintaining human control of LAWS. Endorsing the rolling text would bring Canada into alignment with many close European allies and would contribute to the momentum for a successful conclusion of an agreement on LAWS. To remain on the fence with continued inaction makes a failure of the GGE process more probable, which would likely mean no new global governance structure to constrain the development and deployment of LAWS. Given the strong arguments that fully autonomous LAWS cannot comply with existing IHL in all circumstances — and that existing IHL rules do not adequately constrain LAWS — this lack of governance structure would be inimical to Canada’s stated position as well as its values and interests.
Moreover, even if the CCW process does not ultimately produce a binding agreement, the
substantive work done within the GGE — and Canada’s contributions to it — would not be in vain. Both the Anti-Personnel Mine Ban Convention (1997) and the Convention on Cluster Munitions (2008) had their origins in deliberations within the CCW before negotiations moved to fora outside of it, free of the CCW’s consensus requirement (Herby and Maslen 1998; Grip 2012). Active Canadian engagement and leadership within the GGE now could thus help lay the groundwork for any successor process while also strengthening Canada’s ability to shape its substance and direction.
Support the CAHJC Standard
Canada’s statement within the GGE in November 2025 again called for “ensuring meaningful human control where it is required to uphold those [IHL] principles.”7 The formulation of CAHJC similarly anchors human control in both IHL and context (GGE on Emerging Technologies in the Area of Lethal Autonomous Weapon Systems 2025a). The ICRC’s most recent position paper on LAWS provides further support for the essential elements of CAHJC, stating that the core principles of IHL “presuppose context-specific human judgment” that “cannot be delegated to machine processes” (ICRC 2025, 2). Again, therefore, an explicit endorsement of the CAHJC by Canada would be consistent with Canada’s substantive position and would contribute to the momentum necessary for reaching an agreement within the GGE. The US formulation of “good faith human judgment and care”8 offers no structural guarantee of human involvement or control in the deployment of LAWS and is not consistent with Canada’s long-standing position.
Support the Prohibitions and Situational Regulation
Most importantly, Canada should not only support but seek to take a leading role in negotiations on shaping the prohibitions and regulations for which the language in Box III provides the basis.
Once again, the prohibitions articulated in the tier one paragraphs of Box III are entirely consistent with Canada’s basic position — namely, that LAWS must comply with IHL and must involve
7 See Meeting of the High Contracting Parties to the CCW (2025, para. 5).
8 Supra note 5.
meaningful human control. There is fairly broad agreement within the GGE on these prohibitions as an essential minimal normative floor for any binding agreement. But the ICRC, Human Rights Watch and other organizations have advocated for stronger prohibitions, including specific binding prohibitions on LAWS that target human beings and those that lack meaningful human control (Varella 2025). The first of these prohibitions may be difficult to achieve within the GGE but the second is consistent with Canada’s position.
Where Canada’s engagement would be most significant and meaningful, however, is on the second-tier regulations proposed in Box III to govern non-prohibited LAWS. These measures address precisely the technical and operational risks that make LAWS’ compliance with IHL uncertain, imposing limits on the types of targets that LAWS may engage and limits on the geographic, temporal and situational parameters within which LAWS may operate. The measures require that LAWS operate under a responsible chain of command, with human assessment of legal obligations, and, crucially, they ensure that LAWS mission parameters cannot be significantly modified by the system itself without human intervention.
Canada has credibility on these issues. Its cosponsored working paper on algorithmic bias (GGE in the Area of Lethal Autonomous Weapons Systems 2024a) addressed precisely some of these risks, and its co-sponsored working paper on draft articles on IHL-based prohibitions and regulations (GGE in the Area of Lethal Autonomous Weapons Systems 2024b, 2025b) contributed directly to the architecture of these regulatory proposals. These working papers, coupled with Canada’s longstanding status in disarmament and weaponscontrol negotiations, well position Canada to exercise some leadership in finalizing measures that impose meaningful constraints on the use of LAWS.
Conclusion
Canada’s record on LAWS governance reflects substantive engagement without strategic commitment. It has contributed substantively — on IHL application, human control, algorithmic bias, and the architecture of prohibitions and regulations — but it has declined to commit to the
process that would enshrine these contributions in a new framework to govern LAWS. Negotiations on the rolling text are viewed by many as the last best chance for the development of some global governance structure on LAWS. And even should the CCW process ultimately fail to produce a consensus agreement, it could lay the foundation for a process outside of the CCW, and thus Canada’s efforts within the GGE would both benefit that process and help Canada shape it.
The December 2025 rolling text is consistent with every position Canada has articulated. Canadian support for it would not require any change in substance, but simply the political will to commit to the process before the window of opportunity for governing LAWS closes in November 2026.
AI Disclosure
AI Tools: Perplexity, ChatGPT, Claude, Google NotebookLM, Google Gemini. Conceptualization & Methodology: I used Claude primarily for the purposes of brainstorming the structure of the argument. Information Collection: Perplexity and ChatGPT were used for the purposes of research gathering information and sources on the subject matter. Data Curation, Analysis and Interpretation: Google NotebookLM as well as Claude, Gemini and ChatGPT were used to analyze and summarize sources to assist in the process of synthesizing the information necessary for developing the argument. Writing — Review & Editing: Claude was primarily relied upon for revising and providing editorial suggestions on final drafts of the brief. AI was not used for the purposes of drafting/writing the brief itself.
Acronyms and Abbreviations
AI artificial intelligence
AWS autonomous weapons systems
CAHJC context-appropriate human judgment and control
CCW Convention on Certain Conventional Weapons
DND Department of National Defence
GGE Group of Governmental Experts
ICRC International Committee of the Red Cross
IHL international humanitarian law
LAWS lethal autonomous weapons systems
ML machine learning
Works Cited
Abraham, Yuval. 2024. “‘Lavender’: The AI machine directing Israel’s bombing spree in Gaza.” +972 Magazine, April 3. www.972mag.com/lavender-ai-israeli-army-gaza/.
Bondar, Kateryna. 2025. Ukraine’s Future Vision and Current Capabilities for Waging AI-Enabled Autonomous Warfare. Center for Strategic and International Studies, March 6. www.csis.org/analysis/ukraines-future-vision-and-currentcapabilities-waging-ai-enabled-autonomous-warfare.
Boulanin, Vincent, Laura Bruun and Netta Goussac. 2021. Autonomous Weapon Systems and International Humanitarian Law: Identifying Limits and the Required Type and Degree of Human-Machine Interaction. Stockholm International Peace Research Institute. June. www.sipri.org/sites/ default/files/2021-06/2106_aws_and_ihl_0.pdf.
DND. 2017. Strong, Secure, Engaged: Canada’s Defence Policy. www.canada.ca/en/department-national-defence/ corporate/reports-publications/canada-defence-policy.html.
———. 2024a. Our North, Strong and Free: A Renewed Vision for Canada’s Defence. Ottawa, ON: DND. www.canada.ca/en/department-national-defence/ corporate/reports-publications/north-strong-free-2024.html.
———. 2024b. The Department of National Defence and Canadian Armed Forces Artificial Intelligence Strategy. Ottawa, ON: DND. www.canada.ca/content/dam/dnd-mdn/ documents/reports/ai-ia/dndcaf-ai-strategy.pdf.
Ferrara, Emilio. 2024. “The Butterfly Effect in artificial intelligence systems: Implications for AI bias and fairness.” Machine Learning with Applications 15: 100525. https://doi.org/10.1016/j.mlwa.2024.100525.
GGE on Emerging Technologies in the Area of Lethal Autonomous Weapon Systems. 2024a. “Addressing Bias in Autonomous Weapons. Submitted by Austria, Belgium, Canada, Costa Rica, Germany, Ireland, Luxembourg, Mexico, Panama and Uruguay.” CCW/GGE.1/2024/ WP.5, March 8. https://docs-library.unoda.org/ Convention_on_Certain_Conventional_Weapons_Group_of_Governmental_Experts_on_Lethal_Autonomous_ Weapons_Systems_(2024)/CCW-GGE.1-2024-WP.5.pdf.
———. 2024b. “Draft articles on autonomous weapon systems — prohibitions and other regulatory measures on the basis of international humanitarian law (‘IHL’). Submitted by Australia, Canada, Estonia, Japan, Latvia, Lithuania, Poland, the Republic of Korea, the United Kingdom, and the United States.” CCW/ GGE.1/2024/WP.10, August 26. https://docs-library.unoda. org/Convention_on_Certain_Conventional_Weapons_Group_of_Governmental_Experts_on_Lethal_Autonomous_ Weapons_Systems_(2024)/CCW-GGE.1-2024-WP.10.pdf.
———. 2025a. “Background paper on context-appropriate human judgement and control and the critical functions of LAWS.” CCW/GGE.1/2025/CRP.1, May 20. https://docs-library.unoda.org/Convention_on_Certain_ Conventional_Weapons_-Group_of_Governmental_ Experts_on_Lethal_Autonomous_Weapons_ Systems_(2025)/CCW-GGE-2025-CRP.1.pdf.
———. 2025b. “Draft articles on autonomous weapon systems — prohibitions and other regulatory measures on the basis of international humanitarian law (‘IHL’).” CCW/GGE.1/2025/WP.7, September 3. https://docs-library.unoda.org/Convention_on_Certain_ Conventional_Weapons_-Group_of_Governmental_ Experts_on_Lethal_Autonomous_Weapons_Systems_ (2025)/CCW-GGE.1-2025-WP.7.pdf.
———. 2025c. “GGE on LAW. Rolling text: status date: 18 December 2025.” https://docs-library.unoda.org/ Convention_on_Certain_Conventional_Weapons_Group_of_Governmental_Experts_on_Lethal_Autonomous_ Weapons_Systems_(2026)/CCW_GGE_LAWS_ Rolling_Text_-_status_18_December_2025.pdf.
Grip, Lina. 2012. “Limiting Conventional Arms for Humanitarian Reasons: The Case of Cluster Munitions.” In SIPRI Yearbook 2012: Armaments, Disarmament and International Security, 417–25 Oxford, UK: Oxford University Press.
Herby, Peter and Stuart Maslen. 1998. “An international ban on anti-personnel mines: History and negotiation of the ‘Ottawa treaty.’” International Review of the Red Cross 325. https://international-review.icrc.org/articles/ international-ban-anti-personnel-mineshistory-and-negotiation-ottawa-treaty.
ICRC. 2020. “International Committee of the Red Cross (ICRC) position on autonomous weapon systems: ICRC position and background paper.” International Review of the Red Cross 102 (915): 1335–49. https://doi.org/10.1017/S1816383121000564.
———. 2025. “Autonomous Weapon Systems and International Humanitarian Law: Selected Issues.” December. Geneva, Switzerland: ICRC. www.icrc.org/sites/default/files/2026-03/4896_002_ Autonomous_Weapons_Systems_-_IHL-ICRC.pdf.
Innovation, Science and Economic Development Canada. 2026. Canada’s National Artificial Intelligence Strategy: AI for All. https://ised-isde.canada.ca/site/ised/en/ canadas-national-artificial-intelligence-strategy-ai-all.
Khlaaf, Heidy, Sarah Myers West and Meredith Whittaker. 2024. “Mind the Gap: Foundation Models and the Covert Proliferation of Military Intelligence, Surveillance, and Targeting.” Preprint, arXiv, October 18. https://doi.org/ 10.48550/arXiv.2410.14831.
Kwik, Jonathan. 2024. “Is wearing these sunglasses an attack? Obligations under IHL related to anti-AI countermeasures.” International Review of the Red Cross 926. https://international-review.icrc.org/ articles/is-wearing-these-sunglasses-an-attack-926.
Lee, Major D. A. 2021. “Considering Autonomous Weapons Systems Usage in the Canadian Armed Forces.” Joint Command and Staff Programme 47 Service Paper, Canadian Forces College. https://web.archive.org/web/20240829032036/https:// www.cfc.forces.gc.ca/259/290/23/192/Lee.pdf.
Manson, Katrina. 2026. Project Maven: A Marine Colonel, His Team, and the Dawn of AI Warfare. New York, NY: W. W. Norton and Company.
Marijan, Branka. 2022. “No Canadian leadership on autonomous weapons.” The Ploughshares Monitor 43 (2): 7–9. https://ploughshares.ca/no-canadianleadership-on-autonomous-weapons/.
Martin, Craig. 2025. “Autonomous Weapons Systems and Proportionality: The Need for Regulation.” Case Western Reserve Journal of International Law 57 (1): 255–300. https://scholarlycommons.law.case.edu/jil/vol57/iss1/21/.
Meeting of the High Contracting Parties to the CCW. 2023. “Final Report.” CCW/MSP/2023/7, November 23. https://docs.un.org/en/CCW/MSP/2023/7.
———. 2025. “Working paper. Submitted by Canada.” CCW/MSP/2025/WP.12, November 12. https://docs-library.unoda.org/Convention_on_Certain_ Conventional_Weapons_-Meeting_of_High_Contracting_ Parties_(2025)/CCW-MSP-2025-WP.12.pdf.
Narayanan, Arvind and Sayash Kapoor. 2024. AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t, and How to Tell the Difference. Princeton, NJ: Princeton University Press.
Nature Editorial Board. 2026. “Stop the use of AI in war until laws can be agreed.” Nature 651: 282. https://doi.org/10.1038/d41586-026-00762-y.
Permanent Mission of Australia to the United Nations. 2025. “Informal Consultations on Lethal Autonomous Weapons Systems.” Joint Statement, May 12. https://unny.mission.gov.au/unny/250512_Informal_ Consultations_Lethal_Autonomous_Weapons_Systems.html.
Prime Minister of Canada. 2019. “Minister of Foreign Affairs Mandate Letter.” December 13. www.pm.gc.ca/en/mandate-letters/2019/12/13/ archived-minister-foreign-affairs-mandate-letter.
Scharre, Paul. 2019. Army of None: Autonomous Weapons and the Future of War. New York, NY: W. W. Norton and Company.
Stop Killer Robots. 2025. “Cross-regional group of states ready to start treaty negotiations on autonomous weapons.” September 8. www.stopkillerrobots.org/ news/september-2025-gge-joint-statement/.
United Nations Human Rigths Office of the High Commissioner. 2024. “Gaza: UN experts deplore use of purported AI to commit ‘domicide’ in Gaza, call for reparative approach to rebuilding.” Press release, April 15. www.ohchr.org/ en/press-releases/2024/04/gaza-un-experts-deploreuse-purported-ai-commit-domicide-gaza-call.
United Nations Security Council. 2021. “Letter dated 8 March 2021 from the Panel of Experts on Libya established pursuant to resolution 1973 (2011) addressed to the President of the Security Council.” S/2021/229, March 8. https://docs.un.org/en/s/2021/229.
Varella, Laura. 2025. “Time to Safeguard Humanity from Automated Harm.” Reaching Critical Will CCW Report 13 (1). https://reachingcriticalwill.org/disarmament-fora/ ccw/2025/laws/ccwreport/17313-ccw-report-vol-13-no-1.
———. 2026. “The Final Stretch Before the Finish Line.” Reaching Critical Will CCW Report 14 (2). https://reachingcriticalwill.org/images/documents/ Disarmament-fora/ccw/2026/gge/reports/CCWR14.2.pdf.
Wells, Winthrop. 2026. “Battlefield Evidence in the Age of Artificial Intelligence-Enabled Warfare.” Chicago Journal of International Law 26 (1): Article 11. https://chicagounbound.uchicago.edu/cjil/vol26/iss1/11/.
Winter, Elliot. 2020. “The Compatibility of Autonomous Weapons with the Principle of Distinction in the Law of Armed Conflict.” International and Comparative Law Quarterly 69 (4): 845–76. https://doi.org/10.1017/S0020589320000378.
———. 2022. “The Compatibility of Autonomous Weapons with the Principles of International Humanitarian Law.” Journal of Conflict and Security Law 27 (1): 1–20. https://doi.org/10.1093/jcsl/krac001.
Craig Martin
About CIGI
The Centre for International Governance Innovation (CIGI) is an independent, non-partisan think tank whose peer-reviewed research and trusted analysis influence policy makers to innovate. Our global network of multidisciplinary researchers and strategic partnerships provide policy solutions for the digital era with one goal: to improve people’s lives everywhere. Headquartered in Waterloo, Canada, CIGI has received support from the Government of Canada, the Government of Ontario and founder Jim Balsillie.
À propos du CIGI
Le Centre pour l’innovation dans la gouvernance internationale (CIGI) est un groupe de réflexion indépendant et non partisan dont les recherches évaluées par des pairs et les analyses fiables incitent les décideurs à innover. Grâce à son réseau mondial de chercheurs pluridisciplinaires et de partenariats stratégiques, le CIGI offre des solutions politiques adaptées à l’ère numérique dans le seul but d’améliorer la vie des gens du monde entier. Le CIGI, dont le siège se trouve à Waterloo, au Canada, bénéficie du soutien du gouvernement du Canada, du gouvernement de l’Ontario et de son fondateur, Jim Balsillie.
Credits
Research Director,Digitalization, Security & Democracy Aaron Shull
Director, Programs Dianna H. English
Senior Program Manager Jenny Thiel
Publications Editor Christine Robertson
Graphic Designer Sepideh Shomali
Copyright © 2026 by the Centre for International Governance Innovation
The opinions expressed in this publication are those of the author and do not necessarily reflect the views of the Centre for International Governance Innovation or its Board of Directors.
For publications enquiries, please contact publications@cigionline.org.
The text of this work is licensed under CC BY 4.0. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
For reuse or distribution, please include this copyright notice. This work may contain content (including but not limited to graphics, charts and photographs) used or reproduced under licence or with permission from third parties. Permission to reproduce this content must be obtained from third parties directly. Centre for International Governance Innovation and CIGI are registered trademarks.
67 Erb Street West Waterloo, ON, Canada N2L 6C2 cigionline.org