What has changed most in the WMD landscape over the past 10 years?
Ten years ago, following the invasion of Crimea, the situation was not great, but things are even more complicated now geopolitically. That is primarily due to Russia's invasion of Ukraine and the fallout between Russia and Western states. We have also seen an increase in nuclear rhetoric, which we did not have 10 years ago. Thinking about the bilateral relationship between Russia and the US, we are now in a period where there are no arms control agreements between the two, and they have the biggest nuclear arsenals.
With the expiry of the New START Treaty in February this year, there is now nothing controlling or limiting their arsenals. That does not necessarily mean that we are suddenly in a terrible new situation. But over time, that becomes a problem. It gets into the mindset of thinking, “Are they actually now increasing beyond limits that were in place?” And that increases the security dilemma at a time when trust is already low.
You mentioned the security dilemma. Could you tell us about how that idea of restraint worked during the Cold War? Has that paradigm eroded?
I would say that arms control worked best when both sides independently came to an agreement that was in their interests, militarily and politically: to control, to be controlled, or restrained. In the Cold War period, you had an understanding between the two countries – bilaterally – that arms control was good for both. The problem now is that it is no longer a bilateral U.S.-Russia question. The U.S. is very worried about the increase in China's nuclear arsenal. It is a much more complicated field.
The advantage of the agreements during the Cold War was that they brought consultations, verification, and inspections. Today, that is all lost. This is why everyone is very worried. We are now in a situation where we are losing this culture of having on-site inspections of each other's nuclear sites and exchanges of data. Even though, of course, in the Cold War period the nuclear arsenals were much bigger, the toolkit that they had back then made it work because there was a functioning bilateral relationship.
From what you shared, it seems that back then there was a shared reality, even though the arsenals of both superpowers were larger. Instead, today's arms control discussions seem more complex and uncertain given the advent of AI and emerging technologies. Is that the case?
First, we have to talk about nuclear arms control and emerging technologies together. One of the big fears is the extent to which AI will be integrated into systems of nuclear command, control and communication, so-called NC3. That is a good example of where they are totally intertwined.
Second, AI is changing the nature of warfare. It is changing the way decisions are made on the battlefield because humans are using AI to support decision-making.
Another big area is the speed at which this data is being processed. There are big military advantages to that, but there are also questions about whether decisions are being made too quickly.
Regarding the regulation of AI in the military domain, can you tell us about the REAIM process (summit on Responsible AI in the Military Domain), which you have written about? How does this process support stakeholders in understanding the technology?
Currently, we have a kind of patchwork of initiatives to regulate AI in the military domain. At the United Nations, you have the LAWS (Lethal Autonomous Weapons Systems) process. There is also the informal REAIM process, which was started by the Dutch several years ago. The advantage is that it brings together lots of different stakeholders: governments, both diplomats and military, people from academia, think-tankers, experts, and the private sector. But it has been criticised because it is not fully inclusive. They have not invited everyone. The big absent country from that process has been Russia.
Interestingly, they have now taken the discussions that started in REAIM into the UN General Assembly. This way, every country will be represented, and their vision might be to create a new open-ended working group. In any case, they are also going to continue with these informal REAIM Summits. Kenya has stepped up to host the 4th one in Nairobi in April 2027.
What expertise do you think is important to highlight when it comes to walking into these rooms and sitting at the table?
Even though the world looks very different right now, I think there is still a place for traditional diplomacy and traditional diplomatic skills. All the traditional skills needed to build your relationships with diplomats from other countries are really important. It is not as easy to do that these days. It is not as easy for an American, a British or a French diplomat to build a relationship with their Russian or Chinese counterparts.
With the advent of all this new technology, there is a significant new element that is very challenging for diplomats, who now need to be multi-domain experts. Even if you are talking about nuclear these days, you need to understand the nexus with space security, with AI, with cyber. In a sense, you have to be an all-rounder or, at least, have that within your team.
In your experience, if we focus on disarmament and non-proliferation, what are the skills that you need to make these multilateral negotiations succeed?
You need an issue to be ripe for negotiation, a kind of sweet spot where the big actors are prepared to be legally constrained on something. We do not see that very often at the moment. It is challenging because it requires direction from the very top. Nonetheless, in terms of skills, I would go back to relationship building as one of the most important. Creativity is also important too, because quite often you will be negotiating within the very constrained parameters of your instructions. The ability to think creatively, to build coalitions, if we are talking about multilateral negotiations, is very important in helping to work as a group of countries on solving an issue.
