State Submissions on the Legal Review of AWS and Possible Good Practices
The review of weapons, means, and methods of warfare for compliance with international law has attracted significant attention during the international debate about the regulation and use of autonomous weapon systems (AWS). Legal reviews have been cited as one tool States can use to ensure that any use of AWS complies with their international legal obligations. They were even included in the 2019 Guiding Principles, affirmed by the Group of Governmental Experts on Emerging Technologies in the Area of Lethal Autonomous Weapons Systems (GGE) in 2019. How States view the utility and challenges of these reviews, however, remains subject to wide variances in terms of both substance and practice.
This post explains how the ongoing GGE debate has dealt with the topic of legal review of AWS. It summarises a body of analytical work specifically related to legal reviews of AWS, extracted from submissions made in the GGE and discussed in the Expert Meeting hosted in Sydney in 2023 (reported upon here). This discussion is set to be repeated in April this year, again in Sydney.
The post also addresses how the trends or divergences in this debate might support wider efforts in the development of best practices and information sharing between States, and how legal reviews can serve an important—albeit limited—role in confidence building measures associated with the design, development, and potential use of AWS. This, in turn, holds promise to enhance the legal compliance of AWS, regardless of what the broader regulatory debate about AWS might produce in terms of further regulation or a legally binding instrument.
Submissions to the GGE on Lethal AWS
The conduct of legal reviews of any type is far from universal. Review processes face issues of transparency (as discussed by J.F.R. (Hans) Boddens Hosang here). Further, there is no internationally mandated format or method, with the challenges of autonomy requiring new methods to be adopted by States (see Colonel Copeland, here, on the utility of adopting a functional approach to reviews). That said, certain considerations must necessarily form part of the review process for it to meet a State’s international legal obligations. In particular, a review must assess a weapon’s compliance with applicable arms control treaties and general international humanitarian law rules relating to weapons.
Reviewing all written submissions made by States to the GGE, as well as some selected other participants (such as the International Committee of the Red Cross (ICRC) and the European Union), made from 2013 – 2023, was a useful exercise. It allowed an assessment as to whether, and if, the increased discourse on the utility of legal reviews (as mentioned in Rain Liivoja and Sean Watts’s post, here) has had any real influence on potential legal compliance or regulation of AWS. In collating these submissions, it became evident that since the inception of the debate regarding lethal AWS, there has been little nuance or development in the position of participants in relation to legal reviews, even in relation to what appear to be the established facts of the conduct of legal reviews. There were clear divergences in practice pertaining to the conduct of reviews, most of which related to State practices and the adoption of legal obligations. However, despite of these divergences, the submissions were capable of being distilled into six broad themes, covering key aspects of the conduct of legal reviews as they related to AWS, namely:
1. the nature of the obligation to undertake legal reviews;
2. standards of legal reviews related to AWS;
3. the relationship between legal reviews and other regulatory mechanisms;
4. transparency of reviews and information sharing;
5. modifications to a weapon that necessitate re-review; and
6. independence in the conduct of legal reviews.
When digging deeper into these themes, 18 points describing the practical implications and points for discussion could be further distilled, which were used as the basis for discussion and refinement at the Expert Meeting held by the Australian Defence Force in Sydney between March 28-30, 2023. While this initial report did not seek to engage in the normative debates about the regulation of AWS, it did propose a set of operationalised principles derived from State submissions.
In furtherance of this goal, the initial report was used as a raw facilitation document for discussions at the Sydney Expert Meeting. The result of this Meeting was a refined set of principles appended to the Meeting Report: Enhancing the Legal Review of Autonomous Weapon Systems: report of an Expert Meeting, presented as a list of Elements of Possible Good Practice, expanded based upon input from State representatives, academia and industry to a set of 30. We will return to these elements later in this post.
Themes and Key Challenges Arising from State Submissions
The 2023 Sydney Expert Meeting did not focus on the eventual outcomes of the GGE (whether a legally binding instrument or otherwise), but proceeded on the basis that regardless of what outcome the GGE might produce, there remains a requirement to understand and identify limits on the use of weapons, means, and methods of warfare containing autonomy. In the event a legal binding instrument is adopted, it would be necessary to identify whether a capability was prohibited by that legally binding instrument. The vehicle to undertake that assessment is a form of legal review. Alternatively, in the event that the system containing autonomy is not prohibited outright by international law, there is also a requirement to assess its general compliance with a State’s legal obligations.
Thus, another useful reference point in the discussion on AWS legal review—resulting from the Expert Meeting as also reflected in the Meeting Report—is the identification of key challenges that the presence of autonomy in a weapon system poses for existing methodologies for legal reviews. The discussions in Sydney suggested that these key challenges prompt consideration of alternative approaches in the development or adaptation of States’ legal review processes as they relate to AWS.
The key challenges identified were:
– delimiting the nature and scope of the legal review, considering that the basis for conducting legal reviews differs from State to State, and that autonomy relates to particular functions of a system or system of systems;
– identifying the contextual information needed to conduct a legal review of an AWS, considering that assessing the consistency of autonomous functionality with international humanitarian law may require a broader analysis than is normally undertaken in a legal review;
– establishing the boundaries and interactions between legal reviews and the provision of operational legal advice, given the functioning of the AWS or the environment in which it operates may change during its deployment;
– identifying what might constitute a modification of an AWS that necessitates further review;
– addressing the types of AWS-specific technological issues that might affect the manner in which legal compliance can be assessed during a legal review; and
– addressing persistent challenges in reviewing AWS by examining whether challenges that have been raised in the debate relating to AWS-specific legal reviews are the same as for other weapons, means and methods of warfare.
Possible Good Practices Derived from Existing State Positions
As mentioned above, a refined set of Elements of Possible Good Practices was produced. The list is still nascent, having so far been refined based upon: the practical experience of States that undertake legal reviews, with some having implemented approaches to account for autonomy; the views of academia; input from industry experts; and participation in the Expert Meeting by key organisations like the ICRC. It is hoped that as participation in the drafting process for these elements grows, so too can the potential normative value of these elements.
The next Expert Meeting, set to run again Sydney in April this year, has an agenda set to discuss in further detail what good practices for the legal review of AWS might look like based on this report, with a broader range of State participants, and aims to further refine these elements.
We consider this type of engagement and exchange of ideas can provide a valuable resource to assist in progressing concrete outcomes in the GGE on LAWS. In addition, it provides an opportunity for like-minded States to consider and inform the eventual development of best practices in the legal review of AWS. The conduct of information exchanges, and the potential development of a permanent mechanism to achieve this exchange, will serve to not only further the GGE’s stated information sharing aims but potentially assist in bolstering adherence by those States that are obliged to conduct legal reviews. It could also potentially aid in the adoption of this practice by States that do not yet have legal review processes or are yet to adapt their processes to account for the challenges posed by the introduction of autonomy.
We are certainly not suggesting that legal reviews are a panacea to address all the legal challenges associated with the design and adoption of AWS. However, their use as a confidence building measure, and their universalisation, has great utility as part of a suite of measures to ensure that AWS are designed and adopted in a manner compliant with international law.
***
Lauren Sanders is a Senior Research Fellow at the University of Queensland’s Law and the Future of War Project. This article represents the author’s personal views and does not reflect those of the Australian Defence Force or Australian Government.
Rosie Cavdarski is currently a graduate at law at Allen & Overy’s Sydney office.
Photo credit: Nicolas Halftermeyer