From digital diplomacy to AI-driven diplomacies? Mapping the potential impact of AI on Diplomacy

Introduction

The rapid and consecutive launch of artificial intelligence (AI) applications over the past 3 years has generated a global frenzy. Large language models (LLMs) such as ChatGPT, Gemini, and Copilot have already amassed hundreds of millions of users (Samuel-Azran et al., 2024; Reuters, 2024) and have been hailed as ushering in a new era marked by limitless data-driven insight and new forms of creativity. Global internet users have spent millions of hours interacting with chatbots (Samuel-Azran et al., 2024), which offer companionship, emotional support, and a humorous escape from the malaise of life under late-stage capitalism. Activists, artists, and influencers have swiftly embraced AI image generators such as DALL-E, Luma, or Adobe Firefly to create a stunning array of new visuals (Time, 2024). This global frenzy has catapulted the worth of AI-related companies to the top of global markets, with new companies, such as OpenAI, being valued at $500 billion, and established companies, such as Nvidia, reaching a market valuation of $4.9 trillion. Each new day brings with it a dizzying barrage of new AI startups that promise to revolutionize health care, finance, transportation, education, e-commerce, and legal services. The soaring stock prices of AI companies and the accompanying frenzy may best be described as “The Roaring 2020s.”

The advent of AI has not passed unnoticed by digital diplomacy scholars. The past year has seen a plethora of seminars, workshops, conferences, and academic articles all dedicated to analyzing how AI will impact diplomacy (Manor, 2023; Garcia, 2024; Huang, 2024; Sevin & Eken, 2024). Keynote speakers envision diverse applications ranging from data-driven risk prevention to predictive crisis management. Others have focused on new AI-based cyber threats. Still others imagine that AI-driven diplomacy will enable diplomats and Ministries of Foreign Affairs (MFAs) to analyze vast amounts of online data in near-real time, thus helping diplomats manage change in the international system.

While visions of AI-driven diplomacy are enthralling, they are also highly speculative, as few MFAs have actually begun leveraging AI in sophisticated ways. One way of arriving at a clearer vision of what AI-driven diplomacy might actually look like is by analyzing the history of digital diplomacy scholarship. Although it is brief, having emerged circa 2012 with the advent of social media, studies published over the past decade suggest that diplomats’ use of digital technologies is not merely a matter of digital affordances or diplomatic goals. Studies indicate that diplomats’ adoption of digital technologies is actually shaped by a myriad of institutional, societal, and economic factors (Al-Muftah et al., 2018). To illustrate the possible contours of AI-driven diplomacy, this article begins with an extensive review of the digital diplomacy research corpus. Based on a comprehensive literature analysis, this article identifies six factors that shape diplomats’ adoption of digital technologies to obtain foreign policy goals. Next, using these factors, this article outlines three possible scenarios of AI-driven diplomacy: a “De-evolution scenario,” in which diplomats are reduced to data gatherers, an “Augmentation scenario,” where AI insights enables diplomats to better manage crises, and a “Virtuality scenario,” where AI and virtual reality (VR) facilitate interpersonal communication and negotiations between diplomats. The following section introduces the literature review of the digital diplomacy corpus.

Lessons from the (brief) history of digital diplomacy scholarship

The past two decades have been marked by the rapid digitalization of diplomacy as diplomats and MFAs have increasingly relied on digital technologies to obtain foreign policy goals (Manor & Yarchi, 2026). To date, diplomats have embraced a wide array of digital technologies, ranging from virtual worlds to mass messaging applications, social media platforms, blogs, websites, and virtual meeting platforms such as Zoom (Huang, 2022). Moreover, MFAs have developed their own digital technologies, be it in-house algorithms used to analyze and generate social media content or smartphone applications (Manor, 2019). This process of digitalization has spurred an extensive research corpus that spans hundreds of books, chapters, policy briefs, and academic articles, each investigating how different MFAs employ different technologies.

Notably, the digital diplomacy research corpus suffers from two important limitations. The first is that the vast majority of studies focus only on those technologies that diplomats and MFAs have adopted. Scholars know very little about the technologies that diplomats chose not to adopt and why they did so. Diplomats migrated to social media platforms like Twitter/(X) and Facebook (Mazumdar, 2024), yet they did not migrate to forums and aggregators like Reddit. Diplomats experimented with virtual worlds such as Second Life (Metzgar, 2012; Pamment, 2012) but have rarely leveraged VR and augmented reality. MFAs and embassies are not housed in “smart buildings,” diplomats are not aided by humanoid robots, and very few MFAs employ in-house code writers who author algorithms for data analysis (Manor, 2019). Some of these technologies may be irrelevant to the work of diplomats, while others may be too costly. Yet digital diplomacy scholars have yet to examine why diplomats rejected these technologies, nor have scholars investigated MFA decision-making processes regarding which technologies to adopt and which to reject.

In addition, the digital diplomacy research corpus is marked by frequent comparisons between diplomats and MFAs from different countries. Countless digital diplomacy studies are based on comparing the digital activities of different MFAs and embassies (Manor, 2025; Oloo, 2024; Repnikova & Chen, 2023; Wu & Sevin, 2023) such as contrasting the U.S. State Department’s use of Twitter/(X) with that of the British Foreign Office or the Lithuanian MFA (Manor, 2019). Yet, studies also suggest that each MFA is a world unto itself. MFAs are national institutions, social bodies, and political organs, and each MFA is characterized by distinct features such as communicative cultures, hierarchical relations, professional norms, risk tolerance, and historical roles within governments, which have evolved over centuries (Hocking & Melissen, 2015; Weissman, 2024). It is evident to most digital diplomacy scholars that it would be senseless to compare the use of social media by the State Department and the North Korean MFA, given social, political, and normative differences. But this may also hold true when comparing the use of WhatsApp by British and Lithuanian diplomats. Different MFAs may adopt different technologies in different ways.

Given these two limitations, it may be difficult to envision if and how diplomats will leverage AI to obtain foreign policy goals. Indeed, the adoption of AI may vary from one MFA to another, with some ministries even deciding not to adopt AI at all. One way of overcoming these two limitations is by analyzing the digital diplomacy research corpus and identifying recurring factors that have shaped the process of diplomacy’s digitalization across time and across MFAs. To do so, this article reviewed 90 digital diplomacy articles and book chapters published between 2012 and 2025. This extensive review reveals that scholars have identified six factors that ultimately shape diplomats’ use of digital technologies. Crucially, these factors have been identified in different MFAs and have been found to impact the adoption of different technologies such as websites, blogs, social media, and virtual worlds.

The first factor to shape MFAs’ attitudes toward digital technologies is the metaphors used to describe these technologies. When Facebook and Twitter/(X) were viewed as new town squares, or spaces where connected publics could discuss issues of shared concern, diplomats were encouraged to “engage” with social media users and to create digital outreach teams (Hayden, 2012; Mazumdar, 2024). Yet, once social media were associated with the metaphor of “echo chambers” of hate, diplomats were warned of the dangers of feeding potential trolls (Manor & Kampf, 2022). The media’s framing of digital technologies also shapes diplomats’ use of technology. When social media were framed as the “harbingers of the Arab Spring,” senior officials spoke of the need to converse with global users (Cull, 2013). Yet, once social media were framed as a societal threat due to disinformation, officials demanded that diplomats find ways to debunk disinformation through strategic communications (Bjola & Pamment, 2018). Similarly, when the media framed the Metaverse as the next stage in the internet’s evolution, several MFAs were quick to launch embassies on related platforms. However, as the media’s framing of the Metaverse grew more skeptical, MFAs came to view this technology as costly and ineffective. An important question, therefore, is how does the media currently frame AI and what metaphors are used in these news frames?

The second factor that determines how diplomats use digital technologies is the economic logic of these technologies. The economic logic of social media is based on the constant sharing of personal information, which is converted into data that can be analyzed and monetized thanks to sophisticated algorithms (Zuboff, 2023). This logic demands that social media users lead public lives while constantly sharing personal updates of their successes and failures, as everything once done in private must be done in public and shared online for public consumption (Bauman & Lyon, 2013). The second logic of social media is that individuals must command the attention of their peers in what has been dubbed “The Attention Economy” (Zulli, 2018). Studies have shown that these economic logics drove diplomats to enact performances of transparency online including the live tweeting of diplomatic summits and the live streaming of UN Security Council deliberations (McConnell & Manby, 2024; Strauß et al., 2015). Economic logics also account for MFAs’ online invocation of humor, satire, pop culture references, and incivility, which are all used by diplomats to summon the attention of social media users (Chernobrov, 2022; Cornut et al., 2022; Huang, 2022). As such, assessing the impact of AI on diplomacy must be rooted in an analysis of the economic logics of AIs, which, as will be elaborated in the next sections, is a dual logic of datafication and automation.

The third factor that shapes diplomats’ use of technology is the use of digital technologies by individuals. Once individuals migrated to social media and used it to discuss world events, diplomats also migrated to social media (Seib, 2012). Once individuals congregated in virtual worlds, diplomats launched virtual embassies (Metzgar, 2012). And once individuals sought to learn about the world through diverse news sources including blogs, MFAs launched blogospheres in which diplomats narrated foreign policies and current events (Manor, 2019). Thus, an analysis of how diplomats may leverage AI must also be rooted in an analysis of how AI is already being used by individuals across the world.

A fourth factor is whole-of-government approaches to technology. The Obama administration’s emphasis on using technology for public engagement impacted the State Department’s migration online (Tsvetkova et al., 2020). Yet the opposite is also true, as government bans on technology also shape digital diplomacy. Until recently, Chinese diplomats were absent from Twitter/(X), as Western social media platforms are blocked by the Great Firewall. Moreover, recognizing that MFAs are but one unit within governments highlights the fact that MFAs are part of national security apparatuses, a position that may also impact their use of digital technologies and AI.

Finally, there are several institutional factors that shape diplomacy’s digitalization including the digital zeal of senior policymakers (Hayden, 2012), a risk-prone communicative culture (Hocking & Melissen, 2015), greater autonomy among diplomats who are encouraged to experiment with new technologies, the extent to which higher echelons view digital technologies as a “fad” and a distraction from “proper” diplomatic work, and the median age of diplomats in a specific MFA (Manor & Huang, 2022). Yet perhaps the most important institutional factor is risk tolerance. Digital technologies can prove to be highly disruptive in the sense that they may generate unwanted attention and scrutiny (Bjola & Manor, 2024b). The adoption of social media, for example, was limited in risk-averse MFAs given the fear that tweets and posts may be rejected by digital publics, generating online backlash and negative press coverage (Manor, 2019). This was also the case with the use of provocative content and derogatory language on social media, which was limited in risk-averse MFAs but widely adopted in risk-prone MFAs to the extent that diplomats began using social media to mock and vilify other states (Huang, 2022).

The analysis of the digital diplomacy research corpus indicates that the future use of AI by diplomats may not be uniform and that different modalities may emerge. Yet the analysis also helps identify those factors that could shape the emergence of AI-driven diplomacy. Using these factors, the following section tries to outline the potential use of AI by diplomats while discussing three possible scenarios: De-evolution, Augmentation, and Virtuality.

From AI-driven diplomacy to the datafication of diplomacy

Previous studies suggest that metaphors and media frames have shaped the practice of digital diplomacy. The most dominant metaphors used to describe AI are those that ascribe human agency to non-human machines. As noted in a recent article in Science (Mitchell, 2024), AI systems are called “agents”; they are described as having “knowledge” and “goals” while they are “trained” to “read” and “interpret” data. These metaphors all endow machines with human-like attributes, chief of which is the ability to “reason” and thus provide insights to policymakers.

With regard to media frames, journalists have framed AIs such as ChatGPT or Copilot as incredibly sophisticated and extremely accurate. So sophisticated that newspaper articles frequently discuss ChatGPT’s ability to author college-level articles, to pass medical licensing exams, to score near-perfect grades on entry exams to Harvard Business School, and to author legislation (Kelly, 2023; Rosenblatt, 2023). This, according to some scholars, has led to the “mystification of AI” and the view of ChatGPT as an authoritative source of accurate information (Samuel-Azran et al., 2024). This “mystification” of AI is a major contributor to the AI frenzy described above, as the more accurate an AI, the more it can revolutionize health care, education, transportation, or commerce. Sevin and Eken (2024) add that the “mystification of AI” is increased through the authoritative tone of AI applications. ChatGPT’s or Copilot’s responses to queries bear the imprint of expertise, as every question has a definitive answer written in the language of experts.

The “mystification of AI,” and the emphasis on AI’s sophistication may lead to AI’s adoption by some MFAs, especially those with a risk-prone culture, that allow autonomy and encourage digital experimentation and whose higher echelons view AI as a “game changer” or as adding an important benefit to the work of diplomats. However, as MFAs differ greatly from one another, the future will likely be marked by the emergence of AI-driven diplomacies (plural). Some MFAs, enamored by the promise of automation and facing budget cuts, may adopt AI chatbots that can replace consular officers and automatically reply to comments posted online, as well as debunk false narratives and disinformation spread online. Other MFAs, captivated by the promise of big data analysis, may develop in-house AI applications used to simulate state negotiation tactics as a way of enhancing diplomats’ bargaining skills. Still others, smitten by the notion of risk prevention, may use AI to anticipate international crises by analyzing swarms of online content such as newspaper articles and social media feeds.

Crucially, the aforementioned advent of AI-driven diplomacy may also result in the datafication of diplomacy. While scholars once argued that information is the currency of diplomacy, the use of AI may lead to a shift where data are viewed as the new currency of diplomacy or its most important resource. This becomes apparent when examining the economic logic that underpins AIs, which is the dual logic of datafication and automation. The development of AI applications such as ChatGPT demands training on vast amounts of data. So vast that tech experts are already warning that AIs may soon run out of data, having analyzed all the information available on the internet (Shumailov et al., 2024). If AI is to be used to inform decision-making and policy formulation within MFAs, this will necessitate a constant and steady stream of data. For instance, if an in-house AI application is to be used in simulating negotiations, this AI must be trained on a huge corpus of historic data but also updated with current data on the negotiating tactics of different states. Similarly, an AI system used to forecast future crises would require a constant and steady stream of data including news reports, intelligence analyses, and even social media posts from conflicted regions. Even consular chatbots would require constant data, such as changes to immigration regulations, new travel bans, or specific areas where travel is not allowed or restricted.

The logic of datafication runs the risk of transforming diplomats into “data gatherers.” Diplomats’ primary task may be to continuously gather and feed large quantities of data into AI behemoths, while these AI behemoths will “reason” and “formulate” policy recommendations. Indeed, the promise of AI-driven insight, generated by incredibly accurate and sophisticated AI applications and based on more data than ever possessed by any single diplomat or MFA, may be viewed as a “game changer” by senior policymakers. This is already the case in many other fields such as health care, policing, and city planning, where data-driven insight is seen as the key to successful policymaking. It is here that the metaphors used to describe AI, the media framing of AI, and the economic logic of AI all combine to create a new kind of diplomat, “the data gatherer.” The result would be a sort of de-evolution among diplomats. Like our less evolved ancestors, the hunter-gatherers, diplomats will be less occupied with practicing their craft and more occupied with hunting and gathering scraps of data. This de-evolution scenario raises the question: Where will diplomats find the time to hunt and gather data while busy managing global affairs? The answer may lie in the second economic logic of AI, which is automation. If routine diplomatic functions are automated, then diplomats will have time to commit fully to the task of data gathering and inputting.

The expected automation of diplomacy stems from the economic logic of AI as well as the current use of AI by individuals the world over. Already today, internet users are spending millions of hours conversing with chatbots. What is unique about AI chatbots is that they offer human-like interactions, providing comfort and company. One of the most popular chatbots in the world is “the Psychologist,” an AI-driven chatbot located on the character.ai application. Studies in medical journals suggest that people have meaningful interactions with “the Psychologist” and that they seek its advice on a range of personal issues, in what has been dubbed “Synthetic Intimacy” (Samuel-Azran et al., 2024; Bhat, 2025). Individuals also increasingly use AI to learn about events and actors shaping their world. More and more internet users begin their online journey through AI applications as opposed to webpages like Google, while a growing number of users also query AIs about events and issues making headlines.

This suggests that AI-driven diplomacy will not lead merely to consular chatbots that provide travel warnings but that some MFAs may design AI chatbots able to discuss different issues including daily life, pop culture, and sports, as well as world affairs, while creating emotional ties with users and new forms of Synthetic Intimacy (Bhat, 2025). AI chatbots will lead to the automation of diplomats’ core task—narrating world events and influencing how people make sense of the world around them and a nation’s role in that world. Already now, different AIs created in different countries promote very different world views. The U.S., European, and Chinese AIs offer very different answers to questions such as “Why does the U.S. support Ukraine,” with American AIs highlighting the need to defend Ukrainian democracy, European AIs discussing the Russian threat to the North Atlantic Treaty Organization (NATO), and Chinese AIs discussing U.S. hegemony and its use of Ukraine as a proxy to combat Russia. Diplomatic chatbots of different nations would similarly provide different perspectives on world affairs while subtly shifting users’ beliefs and worldviews through intimate interactions. The automation of diplomacy would likely go beyond chatbots and include the use of AI to generate texts such as press releases, speeches, policy briefs, and summary reports, further increasing the automation of routine diplomatic functions.

Although the vision of data-gathering diplomats seems far-fetched, it is not dissimilar to the de-evolution of diplomats following the adoption of social media. As diplomats and embassies migrated to social media, and as MFAs built social media empires spanning dozens of platforms and accounts (Bjola & Holmes, 2015; Huang, 2022; Mazumdar, 2024), diplomats became content generators. Social media accounts at the individual or institutional level demanded a steady supply of viral content that could summon the attention of average users, journalists, and policymakers (Manor, 2019). Digital diplomacy units were soon staffed with graphic designers and interns tasked not with engaging with digital publics but with authoring a daily barrage of tweets and posts that could be shared across Facebook, Instagram, Twitter/(X), LinkedIn, Flickr, TikTok, and more. The pressure to maintain active profiles demanded that diplomats provide embassies and MFAs with online content, be it a Selfie from a bilateral meeting, an image of an important vote at the United Nations (UN), or a quirky response to a news report. With each new platform, diplomats had to create new forms of content ranging from 280-character tweets to images, videos, reels, and finally, viral dances. While the social media behemoth demanded content, the AI behemoth will demand data.

AI’s impact on diplomacy may thus be threefold. At the institutional level, data may emerge as the new currency of diplomacy and the new basis on which foreign policymaking rests. At the professional level, diplomats’ functions may steadily be relegated to AI applications that automate routine functions. At the individual level, diplomats may transform from foreign policy experts to data gatherers. However, as different societal, economic, and institutional factors can lead to different uses of digital technologies, there is also a second plausible scenario in which AI augments the work of diplomats, as explored next.

Augmented diplomacy in a crisis-prone world

As different combinations of the aforementioned factors lead to different modalities of digital diplomacy, AI may also lead to more highly evolved diplomats who are better able to contend with the sheer speed and intensity with which conflicts and crises now erupt and unfold, as well as the growing number of actors now involved in crises. In this “Augmentation” scenario, diplomacy benefits from data-driven insights, which allow them to exercise their skill with greater precision and efficacy. The “Augmentation” scenario, however, rests on MFAs’ desire and ability to develop in-house AI tools that are able to analyze vast swarms of data in near-real time. Thus, instead of ChatGPT, one may envision a State Department “StateGPT” allowing diplomats to query the vast digital database of the State Department to better manage crises, be it by analyzing past conflicts, summarizing reports and policy papers from conflicted regions, or identifying changes in sentiment and press coverage that may be indicative of crisis escalation in different regions.

The “Augmentation” scenario becomes plausible when one considers the current state of international affairs marked by instability, complexity, and interconnectedness. Some scholars have characterized this state as one of “Permacrisis,” as crises become the new norm in international affairs with the world rapidly transitioning from one crisis to another including the Syrian Civil War, the War in Yemen, the 2014 annexation of Crimea, the Brexit referendum, and the wars in Gaza and Ukraine, as well as heightened tensions in the Indo-Pacific region (Mishra & Sen, 2024). “Permacrisis” also relates to the complex nature of contemporary crises, which see clashes between the interests of numerous actors. The 2024 Israel-Gaza War, for example, was a war between Israel and the Hamas-led Gaza Strip; between Israel and Iran; between Israel and the Lebanon-based Hezbollah group; and between America and Israel, on one hand, and Iranian proxies in Yemen, Iraq, and Syria, on the other hand, and was also another struggle between Russia and the United States and its Western allies. Resolving the Gaza War thus necessitated that diplomats address the interests of many state and non-state actors, each with their own local and regional ambitions.

In “Permacrisis,” regional crises send global ripple effects while global crises manifest regionally, creating a complex cobweb of actors and crises. Such ripple effects are the result of global struggles over hegemony, on one hand, and the growing number of nations that have adopted expansionist foreign policies regionally, on the other hand. These include Azerbaijan, China, India, Iran, Israel, the United States, Turkey, and more. Finally, due to global interconnectedness, a crisis in one system threatens all other systems (Morin, 2008; Morin & Kern, 1999), meaning that financial or health crises can generate military altercations while military altercations can generate financial collapses. Put differently, international affairs are not marked by a single problem that needs to be addressed or a single crisis that needs to be resolved, but by a plurality of crises, antagonists, and problems with which diplomats must contend (Mishra & Sen, 2024).

AI systems may help diplomats to contend with the sheer number of crises now unfolding across the world, as well as the numerous interests that can trigger crises between states including finance, trade, technology, migration, health, and the global and regional aspirations of state and non-state actors. In the “Augmentation” scenario, AI augments diplomats’ ability to practice their craft and to use their expertise to mediate conflicts and address tensions before they erupt. In-house AI systems can help diplomats sift through large volumes of intelligence reports and media stories to more easily grasp how crises are unfolding in near-real time. Similarly, an AI system can analyze past conflicts between states to identify core interests that may bring differing sides to the negotiating table. Sentiment analysis can be used to monitor and analyze leaders’ statements as a means of gauging the risk of escalation in a crisis or as a way of identifying possibilities of de-escalation as leaders signal a willingness to compromise. AI systems may even be used to simulate conflicts between states, offering diplomats new ideas on how to resolve disputes. Yet, in all these cases, AI merely provides diplomats with insight. It is diplomats themselves who ultimately resolve disputes, not bots or AIs that talk to one another. This is thanks to diplomats’ skills such as interpersonal communication, trust building, and the ability to bring conflicted parties together during negotiations.

Crucially, the “Augmentation” scenario is more likely to occur in those MFAs that are risk-prone, that view AI through the metaphor of “data-driven insight” and who value diplomats’ core skills as indispensable interpreters and trust builders.In this scenario, AIs may be used chiefly for analytical purposes, such as allowing diplomats to easily search vast MFA databases, to compare various drafts of accords or agreements while identifying recurring arguments and core interests of states and parties. AIs may also help diplomats test and simulate various solutions to crises, driving creativity. Indeed, a crisis in one part of the world may be solved by the provision of humanitarian aid in another region. In the “Augmentation” scenario, AI does not lead to de-evolution but to the evolution of diplomats, as AIs augment diplomats’ ability to manage crises creatively.

Notably, the use of AI and data-driven insight in policymaking brings to the fore two inherent paradoxes. The first is a temporal paradox in which the future is expected to simply mirror the past. For instance, the use of AI to anticipate another state’s negotiating tactics, or to predict the next diplomatic crises, assumes that future negotiations and crises will be identical to past ones. Yet this is rarely the case. The difficulty of negotiation and crisis management stems exactly from the fact that these are impacted by a plethora of factors and thus constantly take new shapes and new forms (Mishra & Sen, 2024). The crisis caused by the Syrian Civil War is different from the crisis caused by the annexation of Crimea or the crisis caused by Iran’s nuclear ambitions, as each was different in scale, the number of actors involved, the involvement of subnational and transnational actors, as well as underlying tensions and motives. By relying on AI to forecast the future, diplomats may become trapped in a temporal loop, always prepared to contend with the crises of the past while being unprepared to contend with the crises of the future. Notably, this temporal paradox will be less impactful on risk-averse MFAs that may limit their use of AI to routine diplomatic functions such as generating first drafts of press releases that will then be scrutinized by human diplomats. The temporal paradox is most relevant to novel crises but would have less impact on repetitive diplomatic tasks such as offering consular aid or summarizing press reports, as well as repeating crises such as ongoing clashes between Israel and Palestine or Russia and neighboring countries.

The second paradox is the uncertainty paradox, which is also inherent to the use of AI in policymaking, as AIs are always biased. While media depictions tend to suggest that ChatGPT, Claude, and Gemini are incredibly accurate, these AIs are actually riddled with biased data. The reason being that all these AIs are trained on readily available information from the internet including blog posts, social media posts, Wikipedia entries, and the musings of individual users (Manor & Segev, 2025). And if the data on which AIs are trained is biased, then so is their output. For instance, a recent study (Manor & Segev, 2025) found that ChatGPT is heavily biased when depicting nations from the Global South. The study reveals that Global South countries are depicted mostly as suffering from endemic corruption, instability, and violence; as being unsafe to visit; and as offering little contribution to world history and culture as opposed to Global North countries that are shining examples of functioning democracies devoid of violence and dishonest politicians. Recent studies have found consistent biases in popular AIs including racial, gender, and cultural biases (Varsha, 2023).

These biases will also exist in tools used by diplomats, whether these are commercial AIs, government-developed AIs, or MFA-developed AIs. The reason being that such AIs will be trained on data generated by human diplomats who themselves are biased not due to malice but due to human nature. Intelligence reports, summaries of meetings, risk assessments, policy briefs, and all other data generated by diplomats are subjective in nature and may thus be filled with different biases. Therefore, while AI-driven diplomacy may create a sense of certainty as policy is formulated based on objective data and smart machines, it may actually increase uncertainty, as decisions are made based on biased information, leading to unexpected and potentially detrimental results. The source of the bias may, of course, differ. Biased data would hamper diplomats’ ability to choose optimal policies based on past occurrences. Biased models, such as in-house AI models developed by MFAs, would be more problematic as they would lead to errors not just in policymaking but also in all AI-related activities, be it drafting memos and speeches or creating consular chatbots. A third bias may be institutional. Biased institutional uptake, or a biased MFA view of AI outputs as highly authoritative and objective, would skew decision-making by higher echelons who come to overly rely on AI outputs.

Crucially, diplomats may escape these paradoxes via a third possible use of AI in diplomacy, which is actually tied to VR, as explored in the third and final scenario.

Human-centered diplomacy

At a recent conference, Jeffrey Kaplow and Marcus Holmes made an astute observation—that despite the growing use of technology in diplomacy, it remains centered on human interactions. In their paper “Is the Future of Diplomacy Virtual?,” they suggest that technologies that limit human interactions are of limited use to diplomats. Indeed, digital diplomacy studies have found that while diplomats relied on Zoom during the COVID-19 pandemic, many lamented the loss of interpersonal communication, the co-presence of individuals in a negotiating room, and the ability to read body language and facial cues. Although diplomats employed Zoom throughout the pandemic, it never fully replaced physical diplomacy. In fact, studies suggest that Zoom is now part of blended or hybrid diplomacy, in which diplomats establish initial trust physically and only then migrate to virtual meeting platforms (Adler-Nissen & Eggeling, 2022; Bjola & Manor, 2024a). Zoom-based studies are unique in the digital diplomacy corpus, as they offer insight into why diplomats only partially adopted Zoom and why they also partially rejected it. Zoom was of limited use to diplomats because it limited human interaction.

The same is also true of the use of WhatsApp by Ambassadors. Indeed, WhatsApp has not been used to replace human interactions but to enhance them, allowing diplomats to foster closer ties as online discussions continue offline, while offline debates continue online (Manor & Kampf, 2022). This is also true of the use of social media by some states. A recent study has found that some MFAs have remained active on Twitter/(X) not due to their desire to publish tweets but because they use the platform’s instant messaging system to converse with colleagues, peers, and stakeholders. The study elucidates how diplomats are drawn to technologies that facilitate interactions. Since Elon Musk’s acquisition of Twitter/(X), the network has descended into anarchy and toxicity, with fake news, conspiracy theories, trolls, and pornography dominating users’ feeds. This has triggered an “X-odus,” to paraphrase Fishman and Manor (2024), with many users abandoning X. Diplomats in some MFAs considered leaving X, as this digital environment is no longer conducive to interaction and dialogue. Yet they ultimately remained on X to use the instant messaging system and to continue conversations with peers and stakeholders. Here again, one finds that the impulse of diplomats is to become more human and not less human, to increase human interactions and not reduce them by automating diplomacy altogether through bots and AI. AI may allow diplomats to do just that, especially if integrated with VR.

VR allows diplomats to interact closely despite the great distance between them. Like other digital technologies, VR obliterates time and space, allowing two diplomats on either side of the world to feel that they are sharing the same space. In this sense, VR restores co-presence; it enables eye contact, reading facial cues, interpreting body language, and sending non-verbal cues, all of which are essential for building trust, understanding, and fostering ties. In this final scenario, AI may be used to actually enhance human communication, as diplomats in virtual meetings may focus on one another while, in the background, AI systems are used to take notes, record minutes of meetings, produce summaries, and even provide real-time translation. In this scenario, AI does not make diplomacy less human but more human, facilitating a new form of human-centered diplomacy based on combining VR and AI.

VR can drive human-centered diplomacy in three ways. First, it may allow diplomats to share a virtual environment, creating shared experiences and facilitating problem-solving. For instance, diplomats may virtually visit drought-stricken regions or melting glaciers, allowing them to “experience” the impact of climate change. Such a shared experience may bring diplomats closer together while also facilitating a willingness to address shared problems through shared solutions. As described above, AI may play a crucial role in such experiences by allowing diplomats to “stay in the moment,” knowing that verbal exchanges and agreements are being recorded and can later be accessed to begin formulating accords or summaries. Second, VR negotiations may build on AI to help facilitate agreements between diplomats. For instance, two diplomats meeting in a virtual environment may use AIs to generate scenarios, simulations, and models that test the potential impact of different solutions. In this case, diplomats may remain focused on one another and on human interactions while AI systems, operating in the background, help diplomats choose between different options and different paths. Third, AI systems may prove useful in virtual environments by integrating tone analysis and cultural framing, thus reducing miscommunication that can derail diplomatic meetings. Indeed, groups of diplomats may meet for multilateral negotiations in a virtual environment that resembles an offline environment, such as a familiar room at the UN. VR can thus be used to recreate familiar offline settings, allowing diplomats to focus on social cues and body language and maintain eye contact. The AI system operating in the background can facilitate multilateral negotiations by helping diplomats accurately gauge the tone of their peers and interpret their body language and facial cues while also mitigating cultural differences. In all these instances, AI does not reduce human interactions but facilitates human interactions, allowing diplomats from different cultures to better understand one another, making diplomats more human and not less so.

This scenario of AI-driven, VR-enabled human-centered diplomacy is most likely to unfold among those MFAs that belong to governments that have chosen to invest in VR technologies and that view VR as a “game changer” in managing interpersonal communication in a globally interconnected world. Indeed, previous digital diplomacy studies have found that whole-of-government approaches to digital technologies shape the use of technology by diplomats. Moreover, the integration of both VR and AI into the day-to-day operations of MFAs may occur thanks to three institutional factors, namely the digital zeal of senior policymakers who may allocate necessary funds, an institutional mentality that encourages experimentation with new technologies, and greater autonomy for diplomats to integrate new digital technologies into routine functions such as remote negotiations.

It is this third scenario that highlights an important issue: governments’ approach to AI as a whole. Digital diplomacy studies have found that governments’ attitudes toward digital technologies have a direct impact on when and how diplomats leverage digital technologies to obtain foreign policy goals. The U.S. administration, for instance, which prioritizes public engagement, facilitated the State Department’s migration to social media and experimentation with virtual embassies. The Canadian government’s desire to control state messaging directly limited Canadian diplomats’ ability to use social media and to converse with users, stakeholders, and journalists online (Manor, 2019). The adoption of AI may be slower and more limited in scope, in those MFAs that belong to governments that view AI as a national security priority. Here too, governments are not monolithic. Some governments may view AI through the lens of a new “Arms Race” or a tech-based competition over power and hegemony. This is true of certain Western governments such as the United Kingdom and the United States as well as China (Winter-Levy, 2024). Other governments may view AI as a risk to national security given the ability to create new forms of disinformation that warp public opinion and enhance political polarization. Still other governments may be concerned with new and more powerful forms of cyberattacks (Yamin et al., 2021). What all these governments have in common is that AI is subsumed by the logic of securitization, which is at odds with the spirit of digital experimentation. MFAs of such governments may, therefore, be expected to use AI for routine and simple functions such as generating summaries of documents or reviewing newspaper articles. More sophisticated uses, including integrating AI and VR to facilitate remote diplomatic negotiations, will be less tolerated given risk perception.

The opposite, of course, is also true. Human-centered diplomacy, made possible thanks to integrated AI and VR systems, may be eagerly adopted by governments that wish to overcome a small diplomatic presence or who wish to partake in global discussions despite financial limitations. Among such governments, AI may be viewed as a national asset rather than a potential national security threat. The Baltic states, for instance, which have a relatively smaller diplomatic corps, may eagerly adopt integrated VR and AI systems, allowing diplomats in headquarters to remotely participate in various meetings and negotiations. Indeed, Baltic MFAs have proven to be eager adopters of new digital technologies in the past with Lithuania pioneering the use of monitoring teams to combat social media disinformation (Manor, 2019). The same is true of countries in the Global South that wish to engage in greater multilateral diplomacy at low costs through remote participation. It is interesting to note that such patterns of digital adoption have been found in previous instances with Barbados being the first country in the world to launch a virtual embassy in the Metaverse and the Maldives being the second nation to launch an embassy in the virtual world of Second Life (Wyss, 2021).

Conclusions

This article sought to analyze the digital diplomacy research corpus to map out the potential impact of AI on diplomacy and the possible emergence of AI-driven diplomacies. This article posits that AI-driven diplomacy will not be shaped solely by the affordances of different AIs or by diplomatic goals. If history has taught us anything, it is that diplomats’ use of digital technologies is shaped by numerous societal, institutional, and economic factors. As such, there is no single trajectory for the development of AI-driven diplomacy. One plausible scenario is the de-evolution of diplomats who are relegated to data gatherers. Yet a second, and equally plausible, scenario is that AI augments diplomats’ ability to practice their skills of crisis management, while a third scenario suggests that AI will make human diplomats more human and not less so. These different trajectories or visions are each the result of a different combination of factors, ranging from risk tolerance to the economic logic of digital technologies, whole-of-government approaches to technology, the media’s framing of technology, and diplomats’ partial rejection of technologies that limit human interaction.

This article concludes that AI’s impact on diplomacy may be counter-intuitive. Rather than add certainty to diplomacy, AI may increase uncertainty; rather than facilitate crisis management, AI may impede crisis management; and rather than reshape the work of diplomats, AI may usurp diplomats’ role as policymakers, reducing them to data gatherers. However, AI may make diplomacy more human-centric, allowing diplomats to focus on each other while relegating tasks to AIs operating in the background. While the future of AI-driven diplomacy remains uncertain, one thing is certain. As was the case with past innovations, AI will also profoundly impact diplomacy, giving rise to new professional norms, new values, and new working routines, all of which will reshape how diplomats obtain their policy goals.