Dr. Diego Salama, Head of Partnerships and Communications at 糖心Vlog破解版-MERIT, speaks with Dr. Michal Natorski, Assistant Professor, about his recent fellowship at the Dutch Embassy in Seoul. As an Embassy Science Fellow under the , Dr. Natorski conducted research on South Korea’s artificial intelligence strategy, looking at its domestic industrial dimensions, dual-use security implications, and role in shaping global governance. Their conversation touches on the links between AI, diplomacy, innovation, and the future of multilateralism.
Diego Salama (DS): First, tell me what you did in Korea?
Michal Natorski (MN): I went there as a researcher and academic staff member from 糖心Vlog破解版-MERIT. However, I was based at the Dutch Embassy in Seoul as an Embassy Science Fellow, researching a topic defined by the Embassy through an open call for proposals. The competitive call was organised by NWO in collaboration with the Dutch Ministry of Foreign Affairs as a part of the Science Diplomacy Fund (SDF). Academics from Dutch universities go to Dutch embassies, trade offices, and delegations worldwide to perform research that serves the Embassy's policy purposes. It's intended to increase knowledge about topics that the Embassy wants to understand better and enrich academics' understanding of Dutch foreign policy. So, I went to the Republic of Korea to research the country's artificial intelligence (AI) landscape, focusing on both civilian and military uses, in the context of geopolitical and security concerns on the Korean Peninsula and the broader region.
DS: So how does Korea's AI strategy link domestic industrial policy with foreign and security policy: protect the country, but also protect the sector?
MN: Both dimensions are connected, of course. On the one hand, there is domestic development in the AI sector. In fact, AI technology currently constitutes one of Korea's top priorities in the long-term strategy. This is not only declarative, but both the government and private industry conglomerates invest heavily in the various components of the AI stack, from fundamental research and innovation to data centres and semiconductors, as well as data accessibility and the deployment of numerous practical AI tools, including the development of Korean Large Language Models. Korea is well known for its prominent role in digital transformation and the semiconductor industry, which is the backbone of AI systems worldwide. AI is also relevant to the Korean automotive, shipbuilding, healthcare, and electronics industries. The aim is to maintain Korea's role as an exporter of top-quality and competitive goods and services, including those based on AI-enhanced industries. And it also includes creative sectors.
This domestic ambition is also connected to foreign policy. Korea actively engages in multilateral frameworks to contribute to the governance of AI development worldwide. One way of doing so is by organising dedicated summits and events related to AI. My research was also closely related to the REAIM (Responsible AI in Military Domain) summit, which the Republic of Korea organised in Seoul in September 2024 with the Netherlands as a co-host. I could first-hand witness how a summit with over several dozen state representatives and hundreds of stakeholders is prepared.
DS: On the domestic side, which two or three policy instruments are doing the most to drive AI adoption across Korea's public and private sectors?
MN: One distinctive feature of Korea's approach is that there is no single policy instrument—it's a comprehensive strategy that affects the entire society. The Korean AI Strategy was adopted in 2019, but there is space for adaptation according to the needs. This includes substantial investment in science, research, and development for public universities, research centres, and private companies. There's a strong focus on data availability and accessibility, which is necessary to train AI models. Another important element is the opening of data centres. All of these components are interconnected. The broad strategy aims to develop the sector through collaboration between the public entities and companies such as Samsung, LG, SK, NAVER, Kakao, and many others, which are very active in AI research and deployment.
DS: Now, back to the security side of things. Where do you think the dual-use pressure and defence priorities are most visibly shaping civilian decision-making around AI, especially given the security concerns that South Korea faces?
MN: This is one element I think is completely overlooked in international debates. AI is intrinsically dual-use by definition. The same tool can be used for both military and civilian purposes. For example, image recognition can detect enemies on the battlefield, but it can also detect defects in products. So, frequently, the current tendency to distinguish between civilian and military frameworks of AI governance overlooks this inherent overlap.
AI for military and security purposes requires specific developments, but the underlying logic is the same as in the civilian domains. Korea invests in developing AI systems for its military, including the army, navy, and air force. Korea is developing different unmanned systems, intelligent command and control systems, and early warning systems. This research and defence industry effort requires awareness of developments in the civilian sectors. It was not unusual to see military personnel exhibiting prototypes at civilian industry expositions or researchers and private companies presenting their solutions for military use. Given the extremely challenging security landscape on the Korean Peninsula, I think there is a specific awareness in Korea that the military and civilian sectors should collaborate. It also has consequences for foreign and defence policy priorities.
DS: Moving towards the United Nations and the processes you're familiar with, how do you think Korea and the Netherlands fit into global frameworks such as the Global Digital Compact and UNESCO's AI Ethics Recommendations? Are they norm-setters, norm-takers, or hedging actors?
MN: South Korea and the Netherlands collaborate closely within multilateral frameworks. For example, both countries proposed the first UN resolution on AI in the military domain, which introduced this topic to the UN agenda and included states' reports on their approaches to AI in military use. Both countries have also hosted summits on responsible AI in the military domain, the first in the Netherlands in 2023 and the second in South Korea in 2024. These summits have helped develop frameworks and some norms on how AI should be approached in the military and beyond, as well as the complex debates about autonomous weapon systems in Geneva in the framework of the so-called Group of Governmental Experts of the Convention on Conventional Weapons. By the way, a Dutch diplomat also chairs these debates.
However, the REAIM process is much broader. It has been designed to establish a platform for multi-stakeholder debates about AI in the military and as a learning process for involved partners. The Dutch and Korean organisers of these summits were responsible for capturing the consensus views of participating countries in the adopted documents: the Call to Action and Blueprint for Action on responsible use of AI in the military domain. They are declaratory but remain highly relevant as early norm-setting frameworks for approaching AI in the military domain. As a general-purpose technology, AI can have numerous applications in the military domain, ranging from command, surveillance, and reconnaissance to logistics and military healthcare. These are not theoretical debates anymore, but the reality of the war in Ukraine remains with us every day. AI employed in the military can paradoxically save Ukrainian lives and defend against Russian invasion. That is why the Netherlands and Korea promoted a comprehensive debate on these topics. You can imagine that AI-enhanced military capacities can also be an asset for deterrence.
DS: This relationship seems like an important partnership between the Netherlands and South Korea. What are the most actionable Netherlands–Korea opportunities for the next year regarding policy dialogue, R&D, export controls, etc.?
MN: There are plenty of opportunities, and the agenda for governmental consultations between governments is already quite intense. The private sector also plays a major role in AI, and we shouldn't forget that. For example, the Netherlands is closely involved in semiconductor collaboration with Korea, especially through companies like ASML, which provides cutting-edge equipment for Korea's semiconductor industry.
An unexpected observation during my stay at the Embassy was that, besides performing its classic diplomatic roles, Dutch diplomacy also actively shapes the collaboration and partnership between private entities by promoting Dutch innovations. This role of diplomacy is hardly captured in different economic innovation diffusion models. However, it immediately reminded me that classic diplomatic studies do not typically view embassies as intermediaries in innovation diffusion. So, the stay at the Embassy was both a reality check for academic theories and a source of inspiration for future research. I was surprised by the effort put in by the Embassy in stimulating economic collaboration with an emphasis on innovation.
The Embassy has an active agenda for promoting Dutch cutting-edge innovation in South Korea, in digital, quantum, and renewable energy technologies. As a result, there is much potential for economic collaboration, despite the 14-hour flight distance. The broad area of AI is one of the most promising areas for collaboration between the two countries. They have already proven through collaboration in multilateral frameworks that they can contribute to pragmatic solutions in the global AI sector, avoiding the dominance of other powers, such as China and the United States.
DS: What do you think are the domestic and multilateral guardrails that Korea and the Netherlands consider credible for managing risks related to AI escalation or misuse while advancing innovation?
MN: South Korea and the Netherlands are committed to international frameworks like the OECD and UNESCO, which set soft law expectations for appropriate behaviour. At the domestic level, both countries are focused on creating a balance between regulating AI and not stifling innovation. Finding that balance is difficult, but both countries are working hard to achieve it. They focus on ensuring that AI development occurs within a safe regulatory framework while protecting human rights and fundamental freedoms. It is remarkable that the Netherlands, as a Member State of the EU, adopted the first legally binding comprehensive EU AI Act, and a few months later, the Republic of Korea adopted the AI Basic Act, the second in the world to have these characteristics.
DS: I'd like to ask about your experience. You did such an interesting fellowship, and I'm really impressed by your methodology. You ran over 30 interviews, attended five AI expos, and participated in six international conferences. How did you triangulate all these insights from the different venues and meetings?
MN: The goal was to make the research actionable for both policy purposes and academic work. I combined interview insights, written sources, and personal experiences at the AI sector expos and conferences. This in-depth immersion allowed me to interpret the collected data in its specific context.
This experience was unique because I wasn't just reading about the topic but actively engaging with it. Embedding within the Embassy allowed me to interact with people usually inaccessible to external researchers, like government officials and military personnel. I was able to collect insights not just through formal interviews but also through informal interactions. My stay at the Embassy significantly expanded my background knowledge in many ways, as I had the opportunity to meet many people and visit many places. It reminded me that research is not only armchair theorising or crunching numbers, but also that meeting people is still important and inspiring. The research I enjoy is immersion in the field, where I observe the context, including architecture, cuisine, fashion, mobility habits, and everyday life. And it has its rewards. For example, I gained a very clear understanding of data sovereignty by realising that Google Maps does not work for everyday commuting within Seoul. Korea has its own navigation app, Naver, as it is reluctant to share geolocation-sensitive data even with global tech giants because it is also a military asset. Being 50 kilometres from the Northern border, the reason is quite apparent to everyone once you are there.
DS: And the social aspect likely provided additional input. Did informal conversations during coffee or lunch breaks influence your thinking?
MN: Absolutely. The informal social element was very important, but also enjoyable. I was based at an Embassy, which is a very dynamic environment. The Embassy's staff was very welcoming and supportive of my research. There were many opportunities to chat with colleagues over lunch or coffee, and those conversations often provided insights that helped me interpret the more formal data I was collecting. As a person attached to the Embassy, I could participate in official meetings and events.
DS: Finally, multilateralism is in crisis, and the systems are under immense pressure. How has this experience shaped your view on the state of multilateralism today?
MN: Multilateralism has always been stressed, but the current situation highlights the challenges in balancing national interests with global cooperation. The key, however, is that multilateral frameworks provide a space for countries to learn from each other and build trust. The AI in the military domain is a perfect example of this. Despite some countries trying to avoid discussions, there is much interest from other countries in learning about AI's role in the military domain, also debunking the myths and disinformation around it. Multilateral frameworks provide the best opportunity to create trust among actors and share knowledge.
Suggested citation: "Exploring AI, Diplomacy, and Multilateralism in Korea ," 糖心Vlog破解版, 糖心Vlog破解版-MERIT, 2025-09-18, /merit/feature/exploring-ai-diplomacy-and-multilateralism-korea.
