Published in The Ploughshares Monitor Volume 44 Issue 1 Spring 2023
In mid-February, Ploughshares Senior Researcher Dr. Branka Marijan participated in the first global Summit on Responsible Artificial Intelligence in the Military Domain (REAIM 2023) in The Hague. She both created and moderated one of the conference panels. The story of how she came to be there sheds light on the role of Project Ploughshares on the international stage.
Wendy Stocker: Readers of the Monitor are used to reading about Ploughshares staff at meetings of various United Nations groups and gatherings. Over the years, Ploughshares has also been active at meetings related to international treaties, including the Nuclear Non-Proliferation Treaty, the Arms Trade Treaty, and the Treaty on the Prohibition of Nuclear Weapons. But REAIM is a little different and novel and so, we believe, of particular interest to our readers.
Branka, why do you think that Ploughshares was invited to submit a proposal for a panel at REAIM 2023? What do you think the invitation and the acceptance of the proposal says about how the work of Project Ploughshares is viewed?
Branka Marijan: I had written a report on responsible uses of artificial intelligence (AI) by militaries two years ago. Ploughshares has observed the UN discussions on autonomous weapons for some eight years now.
The invitation was sent to organizations and individuals that were known to have followed this issue. I believe that the acceptance of our panel proposal reflects the regard in which our perspective is held. Even those with different points of view like to engage with us.
At Ploughshares, we approach issues with nuance, conduct research with academic rigour, and communicate in policy-oriented ways. This “Ernie Regehr”-style approach, developed by Ploughshares co-founder and first executive director, has remained a key part of our organizational identity. Current executive director Cesar Jaramillo embodies this, as do Senior Researcher Jessica West and Researcher Kelsey Gallagher. I consciously work to develop this quality in my work as well.
If you watch Ploughshares staffers at events and discussions, you’ll see that we adopt a sharing approach. We’re open to new perspectives. We like to dig down and see the truths beneath the surface. We’re not big on ego and we love to collaborate. We build on our Canadian context – we might be the only Canadians in the room!
WS: You were at an event in Ottawa in January that involved the Dutch ambassador to Canada. Describe the event and your role in it, please. How was it connected with the event in The Hague?
BM: I had written a paper for the Centre for International Governance Innovation (CIGI) on autonomous weapons and CIGI invited me to the lunch-time event that it co-organized with the Dutch embassy to highlight some key trends in military uses of AI. I was one of two speakers at the event at the Rideau Club. Dutch Ambassador to Canada Ines Coppoolse gave the opening remarks. I was impressed by her sharp assessment of the role of emerging technologies.
The Ottawa event was one of many organized by the Dutch government ahead of the REAIM summit. I think that the Dutch model is a great example that Canada could use to take on a leadership role in regulating emerging military tech.
WS: I’ve read the Ploughshares annual reports (available on our website) and I know that our budget is finite. I’m guessing that travelling to the Netherlands isn’t cheap. Were you offered any financial support?
BM: We always look for travel support; most of our trips are funded with grants or by event organizers. In this case, the event organizers provided the venue and audio-visual support but did not cover my travel, lodging, or food. However, serendipitously, the focus of this summit related to my research project on emerging technologies and the great power competition, which is supported by a grant from the MINDS program of the Canadian Department of National Defence. The summit helped me to move this project forward by highlighting key issues to explore. And so we were able to use funds from the grant to support my attendance.
WS: Tell me a little about the summit.
BM: The summit was spectacular! Well organized and sophisticated, it brought together leading experts on military applications of AI. Crucially, the summit aimed to bring attention to our concerns about deploying technology that might not yet be ready for the battlefield and, in some cases, should not be used at all. I listened to as many panels as I could, with some sessions held concurrently.
Perhaps my favourite was the discussion on shared challenges in civil and military AI regulation, because clearly AI technologies are multiuse. Defence applications bring heightened concern but there are some shared challenges in the civilian sector that also need to be better understood. Too often the security and defence applications of technologies are set aside as unique cases and not discussed more widely.
My panel was titled “Known Unknowns and Military AI.” It had superb panelists. Elke Schwarz is an Associate Professor of Political Theory at Queen Mary University of London. Lode Dewaegheneire is a military advisor with Mines Action Canada. Arthur Holland Michel is a Senior Fellow with the Carnegie Council for Ethics in International Affairs. I served as participant-moderator, tasked with linking together the points made by the panelists.
I think that we were able to convey clearly the numerous technical, legal, and humanitarian concerns of using AI technologies in defence applications. We stated plainly that even though some of the concerns are already well known, there is always a degree of unpredictability when deploying AI-enabled systems in volatile and dynamic contexts such as war zones.
We did descend into “doom and gloom” territory when talking about lagging regulation and the pace of technological advancement, but I think the audience was very engaged. We used a participatory software, Menti, and had a lively Q&A that went well beyond the dedicated time. We were even praised for organizing a most engaging panel.
WS: What did you take away from the summit? How do you think that your work and the mission of Ploughshares will benefit from this trip?
BM: While I feel that this trip helped to build knowledge and capacity on military AI and will reinvigorate my work, I came away more concerned.
I went in thinking that there was wide agreement on the need to be creative in developing a multilayered governance framework to respond to challenges that AI will bring to warfare. However, all we heard were nice commitments with no specific solutions or roadmap.
The need to rein in and control the development and use of AI by militaries was not universally supported. Some of the scenarios presented in the plenary seemed to indicate that use of AI to carry out tasks such as selecting and engaging targets was inevitable, and even that it was irresponsible not to consider the need for this application.
At the close of the summit, the United States released a framework for a Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy. The Dutch government presented the Call to Action agreed to by a number of states. Both documents feature important commitments but lack substance; the voluntary measures suggested don’t respond to concerns about deployment of potentially harmful technology.
One of the conference rooms used during the summit was called Kilimanjaro. It hosted panels on the need for regulation hosted by Automated Decision Research (ADR), a branch of the Campaign to Stop Killer Robots (to which Ploughshares belongs), and by PAX, the largest peace organization in the Netherlands. Dr. Catherine Connolly, with trademark Irish wit, noted that she hoped that the room’s name was not indicative of the climb ahead of us.
I fear that the climb will indeed be steep and difficult. However, none of us can be discouraged or stop climbing. If we fail to achieve effective regulation of military AI, the consequences could be calamitous.