Oxford University Professor on How Diplomats Are Adopting AI and the Future of Statecraft
Dr. Corneliu Bjola on the Future of StatecraftInterview by Maya Plentz | 16 September 2026Support THE BRIEF. Become a paid subscriber.How Diplomats Are Adopting AI Artificial intelligence is part of the daily work of foreign ministries, but its major impact may not be the automation of diplomatic work. It may be the transformation of the environment in which diplomacy itself is practiced.In my interview with Dr. Corneliu Bjola, Professor of Digital Diplomacy at the University of Oxford and head of the Oxford Digital Diplomacy Research Group, we discussed AI as a tool, and as a medium, and the wider geopolitical environment where these changes in digital diplomacy are happening.“A lot of the discussion is about AI as a tool. But behind this tool there is also a medium—and there is the bigger picture of AI as an environment.”At the level of tools, diplomats already use generative AI for translation, summarization, audience analysis, pattern detection and content production. As a medium, however, AI begins to shape how officials perceive information, formulate messages and engage increasingly segmented publics. At the environmental level, the issue expands beyond software to the infrastructure that makes AI possible: semiconductors, critical raw materials, energy, water and data centres.Listening before broadcastingMuch of the excitement around AI-powered public diplomacy has centred on personalization. Foreign ministries can tailor messages to specific audiences and communicate at far greater speed and scale. Bjola argues that the more valuable capability may be the reverse: listening.AI can help ministries monitor how issues are developing across languages and platforms, identify emerging concerns and distinguish meaningful signals from the flood of online noise. But the system cannot eliminate the need for diplomatic experience. Officials must interrogate where information comes from, how a conclusion was reached and what alternative explanations may fit the evidence.The danger is automation bias: inexperienced users can become impressed by fluent outputs and follow them uncritically. Bjola therefore sees continuous training as essential. The tools used in 2026 are already substantially different from those launched in 2022, and diplomatic education cannot treat AI literacy as a one-off course.AI is lowering the cost of influence operationsThe same systems that can strengthen diplomatic analysis can also make manipulation cheaper. Influence campaigns that once required sizeable teams can increasingly be run by a few people—or potentially one operator working with multiple agents. This changes the economics of disinformation and cyber-enabled influence.A couple of people—one person even—can engage in influence operations on a large scale, using agents. Sometimes for good, sometimes for evil.For foreign ministries, the challenge is to counter a handful of organized adversaries. They may face hundreds or thousands of state, commercial and individual actors operating at machine speed. Institutions built around slow procedures, annual budgets and hierarchical approvals risk responding with what Bjola called “one hand behind your back.”Diplomacy and the imponderablesTo explain what machines cannot resolve, Bjola drew on the findings of the book by Peter C. Grace, The Intelligence Intellectuals: Social Scientists and the Making of the CIA, which explores how elite academics—most notably Yale historian Sherman Kent—transformed reactive gut instinct into systematic, predictive strategic intelligence, one of the architects of modern intelligence analysis. “Imponderables” are factors that resist reliable measurement: political intentions, surprise, adaptation and chance. More data can reduce some uncertainty, but never all of it.This is where Bjola brings in the French philosopher Henri Bergson’s distinction between intelligence and intuition. Analytical intelligence classifies and connects evidence; intuition helps people perceive a dynamic reality while it is still unfolding. In a volatile crisis, diplomats need both.Teaching statecraft on shifting groundBjola has introduced an Oxford course on diplomacy and emerging technologies that covers AI, virtual reality, quantum technologies and synthetic biology. His curriculum separates three speeds of change. Tools such as ChatGPT, Claude, DeepSeek, Mistral and newer agents change fastest. Capabilities—translation, segmentation, pattern recognition and crisis analysis—change more slowly. Strategic problems such as persuasion, deception, signalling, ambiguity and escalation are the most enduring.That structure allows students to examine new systems without mistaking each product release for a new theory of international relations. It also explains why Bjola draws on philosophy, psychology, anthropology, mathematics and literature. Henri Bergson illuminates intuition; Hannah Arendt raises questions about bureaucracy and dehumanization; Ernst Cassirer helps explain how AI models become symbolic filters through which reality and national images are understood.The future diplomatic workforceThe diplomat of the AI era still needs expertise in trade, international law, security, climate or development. But domain expertise must now be combined with technical literacy, source criticism and the ability to work with—and challenge—AI systems. Foreign ministries will need people who understand both policy and the digital environment in which policy is negotiated.Bjola’s message is neither utopian nor alarmist. AI can improve diplomatic listening, intelligence and decision-making. It can also intensify manipulation, fragment the information environment and concentrate power in the companies and states that control the technology stack. The decisive variable is whether diplomatic institutions adapt quickly enough while preserving the human capacity for judgement.Listen or read my interview with Dr. Bjola.Subscribe to our newsletter THE BRIEF, to our YouTube channel, Apple and Spotify Podcasts and leave a review. We would love to hear from you. TranscriptThe following transcript has been slightly edited for clarity.Corneliu Bjola 00:00 Thanks you for inviting me again to the brief, I’m more than happy to discuss with you about new developments in the field, as you suggested.Maya Plentz 00:10 There’s so much happening, in terms of AI, I think when we spoke last with it was before the launch of ChatGPT that made this huge change in terms of how people are using these technologies in diplomacy, and how diplomats are adopting it and, and so on. And I’ve seen that you have been very, of course, following this very closely. Didn’t miss a beat in a sense, because you’ve been always looking into it and writing papers and analyzing and examining all these issues. Your work highlights how the grammar of digital space changes international statecraft. As public diplomacy shifts from basic social media broadcasting to a hyper personalized AI driven engagement, what do you see as the most significant operational gain for foreign ministries trying to project soft power today, and what are the pitfalls?Corneliu Bjola 01:14 Things are moving very fast. That creates a challenge for those both who try to use these tools in various diplomatic areas, but also for those who study how people use AI in the new environment. Starting with the first question you asked about the challenges. When I follow what is happening in terms of this fast acceleration of this digital tools and also the proliferation is actually fragmentation of the AI digital field. I find it quite interesting to go back to some political theories that one of them which I found it particularly inspiring for this moment. It’s a French philosopher, Henri Bergson. And I’m going to explain very briefly why do I think it’s quite relevant in this context to make sense of what is happening.Bergson received the Nobel Prize in 1927 for literature. It’s important because he makes an interesting distinction between intelligence and intuition. With intelligence is what we do all the time. We try to classify things. We try to make connections, try to look at reality and make sense of what is happening. And I think intelligence is quite important in this case. We develop analytical categories for what we see. We try to make a connection between these analytical categories. We try to see how they can be optimized for various functions. But what Bergson insisted is that reality is not static.And we see that happening, especially with the AI in the past two years, three years, quite well. So reality is dynamic, it’s becoming, it’s evolving. So he has this interesting concept, the French concept, duration, that you have to look at how things unfold in particular sequences of time. And I think this duration is interesting because this is where intuition kicks in. It’s trying to understand a little bit, you know, how this process unfolds before our eyes. So intelligence is the analytical part, but then there is this kind of intuitive sense of what we see before we can apply for analytical categories. Intelligence, in many cases, requires confirmation, requires validation, requires empirical evidence. And that takes a little bit of time to make sense, right? When you talk about GPT and agents, to what extent are we going, as OpenAI announced last week, into the age of AGI.Are we there yet? Do we have evidence? So what, on the other hand, if we apply this Bergsonian model, can we have an intuition, whether this is indeed, you know, the right process? So I think this is something that I found particularly interesting to follow. What is intelligence? What is intuition in the way in which we try to understand what is happening? Now, going back to the pitfalls that you mentioned. I think what we see nowadays, again, maybe a little bit of intelligence, maybe a little bit of intuition, is a lot of discussion of AI as a tool.We use it for various things in which we try to apply for translation, try to apply for summarizing text, try to apply for developing clips. It’s a tool. But at the same time, behind this tool, there is also a medium, right? AI as a medium in which you navigate, to swim nowadays in this environment that is created. That’s a different type of challenge.Because when you see a lot of text, you want to be, especially in diplomatic contexts, you want to be authentic, you know, to reflect certain things.So this is not something that, right, it’s a medium that imbues, that defines a way in which this function takes place. And there is also the bigger picture of AI as an environment. We talked about the digital social media was an environment everybody using and creating this kind of, initially it was very optimistic, and then we found our social media has a lot of dark side. And I think now, we also uncover, I mean, we get glimpses, speaking about intuition, we have glimpses of this new environment.Sometimes we have stories about data poisoning, right, in which the data that you receive to make decision making sometimes has been intentionally poisoned, diplomatic, cognitive warfare, this matters a lot. But also we talk about governance.We talk about environment in a bigger, this is another project that I’ve been working on, the diplomatic stack or the technological stack in which environment means not only that icon on your phone, which a GPT or Claude or other other AI tools, but also what lies behind or beneath it, right, you have semiconductors, you have critical raw materials, you have energy, you have water that is necessary for data center, this entire stack that supports the digital infrastructure as they called.So, that’s a new environment and this new environment creates your political challenges. And we already see them, whether it’s the crisis in Denmark, whether, you know, all kinds of negotiations that take place nowadays, in which trade negotiation, which the main aspect is not only tariffs for goods, reduction of tariffs for goods, but also, you know, new dimension that has to do with critical raw materials and other things.So that’s a little bit of a segmentation when you see, we feel that a lot of discussion, I mean, the big pillar is still AI tools. What is this tool doing? What is happening? But there is also a realization that this tool creates their own medium, right, in which you operate. It’s one thing to work with ChatGPT, it’s another thing to work with Claude Claude to just introduce this watermarking, right, that creates a different medium in which you operate.And then there is a bigger environment that these new technologies create and the environment has big, serious geopolitical implications. But what I found most interesting when we tried to zoom in on ministries of foreign affairs, because you mentioned personalization, I think that was, I think, for public diplomacy and those interested in this kind of engagement, the promise. The promise being that you can now tailor your message to the individual, for better or for worse, or micro-targeting type of thing. And in a sense, you can become more efficient in broadcasting messages, in communication, in communicating your intention.But let me reverse that picture. And they say that probably one of the most important aspects nowadays, which is not particularly well discussed, or, you know, people don’t necessarily pay a lot of attention, is the opposite, not to communicate, but to listen. So I suggested that some years ago before AI, with the social media initially, that people jumped first to communicate without understanding, well, first, the audience. And they just hope for the best that we can put a lot of information out there and then, you know, something will work. But what I think is interesting nowadays is this kind of ability to listen to various publics, to various what is happening in real time. And in the time of compression, that can help with the intuition.So when you receive this kind of information about a situation, I don’t know, unfolding, what is noise, what is signal? And that, I think, becomes more efficient for decision-making.Maya Plentz 10:46 You do still need a human behind it. So when they say, oh, okay, we’re going to take away jobs in terms of the foreign ministries, actually, you need more people who are tech savvy to understand how to use these tools well. And we need them to be able to sift exactly through signal and noise, which, of course, these technologies, AI makes it much easier to do, right? But we still need lots of people. We need a workforce that’s well-trained. For instance, I noticed that the Brazilian diplomatic corps is changing. When you look at the CVs of the new hires, the new diplomats that are entering the career, they all have some computer science backgrounds, for instance.Corneliu Bjola 11:42 I think this is a must nowadays. I think we’ve been talking about this for a number of years now, that the traditional professional profile for entering ministries of foreign affairs has changed. Yes, you need a degree in economics or political science or international relations to be able to have that type of background knowledge to make sense of different things. But at the same time, you operate, speaking about medium, you operate in a totally different environment.The environment is not only analog. It’s very digital. So if you don’t understand this, it’s very difficult then to move very fast. In this kind of context, you’re going to lose opportunities for intervention in various discussions, in various debates.Keep in mind also that nowadays there is a lot of discussion about agreements on governance, on these technologies. So you need to have some technical background. But that applies more generally to the digital sphere.With AI, it’s a new environment. And this is where I think we are scratching the surface nowadays. People are very excited about AI. There have been a number of studies published in the past two years, suggesting that, and actually some of the studies are quite alarming, suggesting that people who have minimal exposure to AI get very excited about it and then they uncritically follow whatever AI tool is telling them.So you need, I mean, there was an interesting, I think in ISQ published an article with people from national security apparatus, that you need a little six months of training, but training is not static because these tools evolve in order to be able to use these tools with confidence, not in a blind way, in a blind manner, following uncritically what is, but also probing, asking questions.Where are these data coming from? How did you reach this conclusion? Can you offer me some other thoughts about this? In a sense, it’s a very dynamic way in which we interact with the tools. And that requires different type of skills that needs to be updated on a regular basis because these tools change very quickly.The AI that’s used nowadays is different from the one in 2022. I mean, it’s quite ancient, that one. So this is what I’m saying here, in a sense, you know, that yes, the humans are still important, are very important, actually. But also in terms of the type of emphasis that probably we need to look, it’s not only communication, it’s also the basic listening, understanding, and listening is important because it develops going back to Bergson’s that type of intuition. When you see certain things, AI can make some suggestions, but does it feel right? And that comes with experience, comes with being immersed in these kinds of environments.So exactly that point about separating white noise from a signal, which intuitively you may develop some hypothesis, transform it to intelligence, but it is important to make connections, connections from different things that are not necessarily quite obvious. AI can help a lot mapping, but your experience as a human operator could make a huge difference if you know how these things, you know, if you are constantly adapting to the new functions of these tools. So I think, you know, this is one, probably one aspect that might be relevant in terms of digital influence. Again, this is, to a certain extent, basic stuff. Don’t jump directly to communicating before you understand, well, who are you talking to, and what seems to matter for that type of audience.Maya Plentz 16:14 And in terms of cybersecurity fractures and open source risks, the cybersecurity landscape is shifting rapidly, as you said, open source repositories like Hugging Face and the deployment of advanced models by OpenAI have lowered the barrier to entry for highly sophisticated, automated influence operations. Given that bad actors can deploy customized large-language models for localized disinformation campaigns at a fraction of the cost, how should ministries of foreign affairs adapt to their defensive public diplomacy strategies to counter these decentralized tech-driven threats?Corneliu Bjola 17:00 That’s interesting, because what is happening with these tools? And I think it’s changing a little bit the economics in which this communication and influence activity takes place. Remember, in the past, if you wanted to do, you know, a public diplomacy campaign, you required a team, a large team, to understand what is happening, to promote, to engage with actors. What happens nowadays with AI, it collapses that type of chain and reduces tremendously the cost.There is a lot of, just as a parallel example, there’s a lot of talk nowadays in business, about creating one-person companies with AI, companies that you can run many things because you have agents and so on. And they think the same thing applies here. So you can reduce significantly the size of the operation. A couple of people, one person even, can engage in influence operation on a large scale, using agents, using all kinds of, sometimes for good, sometimes for worse.And I think for ministry of foreign affairs, that’s a big challenge. How to operate in an environment in which you don’t deal with a thousand of other actors, you can deal with hundreds of thousands, because it’s a different economics that inform this kind of dynamic. So that’s one thing that we see nowadays on cybersecurity, that’s even worse, because on cybersecurity we used to have North Korea-style operations with many institutions operating or Russian operations.The cost of doing that decreases significantly. So you may have state actors, but also non-state actors, companies, but also individuals engaging in this kind of cyber attacks or cyber activity quite successfully, given the new tools.And that challenges many institutions, which operate on a different, probably a different format, still with more and more institutions. So I keep saying that institutions have to become more agile for this new environment.You have to adapt very quickly to keep an eye on the horizon, what seems to be the trend nowadays, and how to make sure that institution is fit for purpose. And that’s not easy. Institutions that we have generally are designed for a different age, with procedures, with budgets that operate on a different scale.So when you have this kind of operation, you will, in a sense, operate with the hand behind your back, because you have to follow a different tempo in order to react to certain events. So that’s one challenge.I think it’s important to see how this decentralization that you mentioned can go very micro and create a plethora, you know, a large number of countries. So actors that are involved in this. There is something else that probably is important here when we talk about influence operation, the decentralized influence operation. And this is, again, something that requires skills and training. In a sense, it’s not quite new. There is an interesting book that has been just published.In this book that was published by Georgetown University by Peter C. Grace’s The Intelligence Intellectuals, it’s mainly about how the CIA was set up in late 40s, a number of people there were involved and tried to think about how this organization should operate in peacetime.The story is interesting, but there is an interesting concept that is covered by one of the CIA founders, the idea of imponderables. Imponderables is an interesting concept. I think Sherman Kent was the one who proposed it in this book. But we have others at that time who posited this idea, George Kennan also used that in a different context.Imponderables refer to things that cannot be measured very well. Political intentions of other parties, surprises, adaptation, various chances. So what happens is that the imponderables, they say, are of two kinds.Some imponderables that you can reduce the uncertainty around them because you accumulate, you acquire information, and there you can help with this. But there is still a residue, something that you’ll never be able to cover well, even with more intelligence, because the situation is moving very fast. And this is, again, Bergson, intuition comes into play.In addition to the fragmentation and decentralization, I think we have also to understand how to, in this avalanche of information, how to deal with imponderables.Imponderables that can be, information can reduce and transform them into something that is actionable. But there is a huge, it’s not quite a residue, there’s huge amount of this imponderables because of, I mean, CIA in 1947 operated with three, four memos per day, let’s say. Today, at this moment you operate with a huge amount of information, some of them reducible, it can be classified and categorized and transformed into analytical insight.But a lot of it, it remains this kind of, unclear because things are very dynamic and very much changing. So I think this is a skill to be developed here, speaking about the medium, speaking about the way in which you operate. I mean, this is an opportunity for diplomacy to become smarter. Not replacing human with the machine, but the machine can help, but the human has to become smarter because of the machine and be able to do certain thing that were not available to that person before. So instead of going lazy, there is a danger of going lazy and, complacent because these tools can do a lot of redundant work. The general for ministries of foreign affairs is how to handle new challenges in a smarter way.And one aspect which I mentioned here is how to cover this situation of imponderables. Peter C. Grace, in his book, made that quite interesting for me with the new tools, and that provides better intelligence, better way to understanding the reality. And there is a mixture of intuition and at some point, you know, intelligence in terms of analytical insight. It’s an exciting period. I think if we don’t miss these opportunities, if you don’t go complacent and hope the machine will do the work for you. It’s an opportunity for those working in this area to become even more professional and more sophisticated in the way. And through their advice to improve decision making, which is much needed given the turbulence and the chaos that is unfolding geopolitically on a daily basis.Maya Plentz 25:43 So, yes, indeed. In fact, it seems like really we need diplomats to be as good as they were before in their respective areas of specialization, be it intellectual property law, trade, or environmental law or what not, but that they are able not to just outsource much of their thinking and analysis to these tools, but actually be very to the point and updated in what the field is going through and then using these tools as a partner, right, to deploy intelligence in, it seems to me.But now I’d like to speak a little bit about pedagogy and the speed of technological disruption. The velocity of AI in cyber security is staggering with newer vulnerabilities and model capabilities emerging with it. How do you structurally adapt your syllabus at the University of Oxford to ensure that graduate students are learning enduring theoretical frameworks rather than fleeting technical cases studies? In other words, how do you teach the future of statecraft when the technical ground shifts beneath the curriculum every semester?Corneliu Bjola 27:10 This is a very timely question because I just introduced a new option this year, which is about diplomacy and emerging technologies. And emerging technologies cover not only AI, cover virtual reality, it covers also quantum technology, synthetic biology, and whatever may come next. But it’s a challenge indeed, as you pointed out, how to think about this issue in a way that can train the minds of young students, of new students about these new challenges from an international relations, diplomatic perspective.There are two general principles that I follow. One is a question of, as you mentioned, the question of acceleration. I try to divide things in three pieces. So one is about the tools. The tools change very fast. The tools are the one that we are familiar with. Open AI, chat GPT, the new one, the Chinese version, and so on. They change every three, four months. There is a new model with new functions, with new attributes. And we have to keep that in mind.There’s another thing which changes a little bit slower. It’s about capabilities. So how these tools are applied to different things. We started with translation, with image transformation, with audience segmentation, large-scale pattern detection. So these are capabilities that you have to understand how these tools apply to them in what way. And these capabilities, some of them novel, some of them come from some previous forms of engagement. They move a little bit slower.And then there is something that changes even slower, which are the strategic problems, right? When we talk about persuasion, rights of power. When we talk about deception, influence operation. When we talk about signaling, ambiguity, escalation, crisis, these are strategic problems.So you have to look at how these different three components in terms of acceleration changes, right? So the tools are the ones that move faster. They may form different capabilities. Capabilities have their own dynamic and, you know, composition and the way in which it operates. And of course, all comes together to source strategic issues. So you look at them at the compact and the feedback loop that exists between them. So that’s one thing, in terms of trying to follow how everything fits together.The second principle, which I found particularly important nowadays, is to go interdisciplinary. And interdisciplinary means that you have to draw on readings from different fields. I mentioned Bergson because philosophy is quite important. It’s not only for me, it also we see nowadays big companies hiring, recruiting philosophers to tell them a little bit more to make sense of what does it mean, this technical tool that they have developed. I think Claude had hired some philosophy. Philosophy for each shift. So I’m trying to find good thinkers. I was designing the core syllabus a few days ago and a question of ethics of AI rose.I went back to Hannah Arendt. And Hannah Arendt comes from a different age, but she speaks about the role of bureaucracies in which one becomes a cog. You become a cog in a machine and you become dehumanized. And there is that type of thing that doesn’t apply completely, but imagine if you are becoming surrounded by all kinds of AI agents in an institution. I think, by the way, when you talk with business, that’s a new big challenge nowadays.They don’t know what the AI agents are doing in their company. They lost track. There are so many nowadays doing different tasks and all of a sudden, and then, you know, the human sometimes, gets some input from them. So the environment is speaking about the way in which you operate has changed. You have to be alert about certain things and how your decision as a human can have probably not as catastrophic as it described by Hannah Arendt, but it has this kind of sensitivity about you operating to an environment that becomes very impersonal.So what I’m saying here is that, yes, you have to pay attention how different important components connect together, the tools, the capabilities, the strategic problems, but also to bring into the fold thinking. Thinking coming from philosophy, thinking coming from literature, thinking coming from mathematics. So I’m not one reading that I found particularly interesting to include in a specific session.Whether it’s about decision making, you have to go back to understand, you know, how human takes decisions, what is the role of emotions, a little bit of psychology. Psychology is a big field nowadays. When we talk about the dark side in the cognitive warfare, this is a totally new discipline that is being developed nowadays.Psychology is important one, but questions of anthropology, you may find quite relevant as well in this context to understand the symbol is the way in which various elements come together. In the article they published recently with a colleague, Ilan Manor, we talked about, we looked at different AI tools, ChatGPT, Claude, and Deepseek, and Mistral, asking similar, the exact 10 questions, all three of them, and we got different answers. It was interesting because this is a new symbolic projection that is being created by these tools.So when you get an answer from Claude, or from Deepseek, or from another, it’s not the same thing. It has nuances embedded into it. And this is where, for instance, we refer to Ernst Cassirer, a philosopher, a German philosopher, who wrote a lot about symbolic forms, symbolic forms being these kind of areas in which, that we use, lenses that we use, lenses that we use to make sense of reality. He wrote it for a different age, but they become quite relevant when we think about how the role of AI in developing these kind of new realities and also helping us understand these new realities.Reality is mediated by these new powerful tools. So this is what I think pedagogically, for the time being, I found quite exciting, how to go from micro to macro, from the tools to capabilities to the enduring strategic problems, but also how to bring insight from various disciplines, different perspectives.Having different perspectives helps. You can go technical, you can understand this, you can go very instrumental, how this, you know, can improve translation or summarizing text or crisis management. But you don’t get the right meaning that being attached to this. I think probably there’s a more exciting aspect with these technologies, both for those who practice, you know, who use these kind of tools in the diplomatic activity, but also for us, like me, who study how it’s being used.THE BRIEFSources and further readingThe three sources Dr. Bjola mentioned during our conversation:1. Henri Bergson: intuition, intelligence and duration; The most relevant primary text is Bergson, H. (1912). An Introduction to Metaphysics. Translated by T. E. Hulme. New York: G. P. Putnam’s Sons; key source for Bergson’s distinctions among intelligence, instinct and intuition, and for the argument that intelligence spatialises, divides and treats reality as manipulable objects.2. Peter C. Grace, The Intelligence Intellectuals: Social Scientists and the Making of the CIA (Washington, DC: Georgetown University Press, 2026); Grace examines how William Langer, Sherman Kent, Max Millikan and other social scientists established the CIA’s early culture of systematic estimative analysis. It is the relevant source for connecting the imponderables, institutionalised judgement under uncertainty. 3. Corneliu Bjola and Ilan Manor, “Encoding Geopolitical Influence: Soft Power and the Symbolic Politics of Artificial Intelligence,” in Naren Chitty, Craig Hayden, Hendrik W. Ohnesorge and Chenjun Wang, eds., The Anthem Handbook of Soft Power and Public Diplomacy in the Age of AI(London: Anthem Press, 2026), pp. 161–178.The main Cassirer sources behind our article are:Ernst Cassirer, An Essay on Man: An Introduction to a Philosophy of Human Culture (New Haven, CT: Yale University Press, 1944) - The most concise source for the idea of humans as animal symbolicum and for language, myth, art and knowledge as symbolic forms.Ernst Cassirer, The Philosophy of Symbolic Forms, 3 vols., trans. Ralph Manheim (New Haven, CT: Yale University Press, 1953–1957 [1923–1929]).- The fuller philosophical foundation for understanding media and AI as structures that organise meaning rather than merely transmit information.• Oxford profile for Corneliu Bjola• Digital diplomacy in the age of technological acceleration• First contact: integrating generative AI into digital diplomatic intelligence• Encoding Geopolitical Influence in The Anthem Handbook of Soft Power and Public Diplomacy• Diplomacy in the Age of Artificial Intelligence• The Intelligence Intellectuals by Peter C. GraceDigital diplomacy in the age of technological accelerationWritten with Ilan Manor and published in Place Branding and Public Diplomacy, this article develops three scenarios for generative AI’s impact on digital diplomacy. Its most useful contribution is to reject a single deterministic future. Generative AI may enhance existing practice, alter diplomatic routines more deeply, or drive broader disruption depending on institutional choices, professional norms and technological development. The paper provides the conceptual bridge to Bjola’s interview distinction among tools, capabilities and enduring strategic problems.First contact and digital diplomatic intelligenceBjola and Manor examine how generative AI can be integrated into digital diplomatic intelligence: the collection and interpretation of online information for foreign-policy decision-making. The attraction is clear—rapid synthesis, pattern recognition and the ability to process volumes of material beyond human scale. The article is equally concerned with technical and organizational barriers, including unreliable outputs, biased data, verification, institutional trust and the need to redesign workflows rather than simply bolt a chatbot onto existing procedures.Encoding geopolitical influenceTheir newest AI-focused chapter, “Encoding Geopolitical Influence: Soft Power and the Symbolic Politics of Artificial Intelligence,” shifts the analysis from efficiency to meaning. Comparing leading systems such as ChatGPT, Claude, DeepSeek and Mistral, the authors ask how models encode different cultural and political assumptions. Drawing on Ernst Cassirer’s philosophy of symbolic forms, they treat AI systems as new gatekeepers through which users encounter nations, identities and geopolitical narratives. In practical terms, soft power is no longer projected only through speeches, media or cultural institutions; it can also be embedded in the defaults and answers of globally used models.Diplomacy in the age of artificial intelligenceBjola’s 2019 paper remains the foundation of the later work. It separated AI as a diplomatic instrument from AI as a force reshaping the environment in which diplomacy takes place. The interview updates that framework for the generative-AI era: the instrument is now widely accessible, the medium is becoming personalized and synthetic, and the environment has expanded into a geopolitical technology stack involving compute, chips, minerals, energy, water and corporate power.Across these works, Bjola’s research has moved from functional adoption toward a richer theory of AI-mediated statecraft. Its strongest feature is the insistence that technological capability cannot be separated from institutions, professional judgement, symbolic power and material infrastructure. A limitation—also acknowledged implicitly in the interview—is that the technology changes faster than conventional empirical research can validate claims. Bjola’s answer is not to abandon evidence, but to pair analysis with cultivated intuition and interdisciplinary thinking. This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit theunbrief.substack.com/subscribe
