AI and International Politics, Economics, and Missions

Lee Cheol-young

(Director, International Affairs Research Group)

Abstract

Drawing upon Yuval Harari's Nexus, Henry Kissinger's The Age of AI, and Daron Acemoglu's Power and Progress, this paper examines, from an integrated political, security, and economic perspective, the macro-level implications that the paradigm shift brought about by artificial intelligence (AI) will have on the future missionary environment. Politically, it warns that AI's opaque black-box algorithms and comprehensive data monopolization may paralyze the distributed self-correcting mechanisms of democracy and usher in a powerful form of totalitarianism. From the perspective of international security, it argues that the introduction of cyber and AI weapons possessing autonomy and non-human logic could undermine existing balances of power and trigger unpredictable, ultra-rapid outbreaks of war and military escalation. Economically, it forecasts that AI may fail to produce universal prosperity and instead remain a form of "so-so automation" that primarily replaces low-skilled labor, shrinking employment opportunities. Furthermore, in developing countries, AI may function as an "inappropriate technology" that removes the ladder of economic development, thereby reinforcing global inequality.

These structural crises will compel a comprehensive restructuring of future missionary paradigms across short-, medium-, and long-term horizons. As AI surveillance systems become increasingly sophisticated, obtaining visas and maintaining covert identities will become more difficult. Consequently, the traditional model of long-term settlement ministry in a single region is likely to decline, while a Paul’s model of highly skilled, long-term itinerant missionaries will become more widespread. As the effectiveness of early-warning systems for military conflict deteriorates, face-to-face ministry in conflict zones may become impossible, leading to the rapid rise of remote missions utilizing digital infrastructure and media technologies. In response to the disappearance of low-skilled jobs and the increase of economic refugees in mission fields, the need will grow for a transition from traditional profit-centered BAM (Business as Mission) models toward BFM (Business For Mission) approaches that support appropriate-technology entrepreneurship and AI literacy education. Ultimately, the missionary community in the age of AI must courageously abandon non-essential pursuits such as methodological sophistication and efficiency, which are vulnerable to technological manipulation and distortion. Instead, it must return to the essence of humanity and the Gospel, focusing wholeheartedly on the Father's heart for lost souls. Only through such a return can the Church overcome the coming challenges.

1. Introduction

The academic concept of Artificial Intelligence (AI) was first publicly discussed in the summer of 1956 at a conference held at Dartmouth College under the sponsorship of the Rockefeller Foundation. What was then regarded as little more than an intellectual curiosity or a speculative concept has, in contemporary society, become an indispensable technological paradigm that shapes human life itself. According to research conducted by the Microsoft AI Economy Institute, as of March 2026, approximately 17.8% of the world's working-age population (ages 15–64) is already utilizing commercialized generative AI products. Among the twenty-six countries with the highest rates of technological adoption, that figure approaches 30 percent. The spread and acceleration of AI technology are therefore ongoing realities, yet academic predictions concerning the future landscape that AI will create remain sharply divided.

Representing a classic example of technological optimism, futurist Ray Kurzweil, in his book The Singularity Is Nearer, predicts that around 2029 AI will pass the Turing Test, becoming capable of conducting conversations indistinguishable from those of human beings. Within academic circles, this stage is generally regarded as the arrival of Artificial General Intelligence (AGI)—a form of intelligence capable of matching or exceeding human cognitive abilities across a broad range of domains. Kurzweil argues that the emergence of AGI will not only produce explosive increases in productivity but will also lead humanity toward a technological singularity by 2045. At that point, he predicts, the human neocortex will be integrated with cloud networks, enabling humanity to overcome biological limitations and achieve immortality. In contrast, Daron Acemoglu, the MIT economist discussed later in this paper, presents a fundamentally different perspective. Drawing upon empirical data and the structural limitations of current AI development, Acemoglu expresses deep skepticism regarding whether AI can truly generate long-term and transformative improvements in humanity's economic well-being.

Thus, intense debate continues within academia regarding the future that AI will bring. This paper focuses on the macro-level effects that this technological paradigm shift will have on future missionary environments. Specifically, it examines the perspectives of leading scholars through the lenses of three key dimensions that significantly influence modern missions: politics, international security, and economics. Based on these analyses, the paper seeks to explore practical directions for future missionary strategies. More specifically, in the political sphere, this study analyzes Nexus, the recent work of Yuval Noah Harari, professor at the Hebrew University, which examines the crisis and future of modern democracy. In the field of international security, it reviews The Age of AI, the final major work of Henry Kissinger, Nobel Peace Prize laureate and architect of modern shuttle diplomacy who helped broker peace between Israel and Egypt. Finally, in the economic sphere, it considers Power and Progress by Daron Acemoglu, recipient of the 2024 Nobel Prize in Economics and a leading authority in the economics of AI. By organically connecting and analyzing the central arguments of these three works, this study seeks to identify the challenges that future missions will face amid rapidly changing global conditions and to establish a missiological foundation for responding proactively to those challenges.

2. AI, Information, Democracy, and Totalitarianism – Centered on Yuval Harari’s Nexus

2.1 Understanding Information: From “Representation” to “Connection”

In understanding human history, the concept of “information” has long been regarded as an instrumental tool for accurately describing reality or conveying truth. We have commonly maintained a “naive view of information,” believing that the free flow of information naturally represents truth, and that the most effective solution to misinformation or disinformation is simply to supply more information to the marketplace. However, Yuval Harari directly challenges this optimism, arguing that the defining characteristic of information is not representation but connection. Information is a social link—a nexus—that connects different points into networks. The purpose of these connections is not primarily to discover truth, but to create social cohesion and enable large-scale cooperation.

The birth of human civilization was made possible through the combination of two information-processing tools: bureaucracy and myth. Bureaucracy enabled the management of large societies through the tedious tasks of collecting, classifying, and visualizing data. Myths created “intersubjective realities” such as gods, money, and nations, allowing tens of thousands of strangers to unite as a single community. What is important here is that intersubjective realities do not necessarily rest upon objective truth; rather, they exist within the stories people exchange with one another. Thus, the history of information is essentially the history of how these networks of connection have evolved. Harari argues that the true protagonist of human history has never been Homo sapiens but rather the flow of information itself. Furthermore, contemporary science increasingly redefines biology, politics, and economics through the lens of information flows.

2.2 Anatomy of Dictatorial Information Networks: The Illusion of Infallibility

2.2.1 Centralization and the Absence of Self-Correcting Mechanisms

Harari reinterprets democracy and dictatorship as contrasting forms of information networks. The defining characteristic of a dictatorial information network is extreme centralization. All information flows converge toward the apex of power, where the assumption prevails that “the center cannot be wrong.” Under dictatorship, independent centers of authority—such as an autonomous judiciary, free press, and academic communities—are viewed as threats to the regime and are systematically weakened or eliminated. The fatal flaw of such a structure is the absence of a self-correcting mechanism. A dictatorship is a closed network lacking robust feedback loops. When errors occur at the center, there is no effective means of correcting them, creating a constant risk of catastrophic outcomes.

An important distinction must be made between dictatorship and totalitarianism. Not every dictatorship is totalitarian. Totalitarianism refers to the extreme attempt by the state to exercise complete control over every aspect of citizens’ lives, as exemplified by Hitler and Stalin. Although many dictators throughout history harbored totalitarian ambitions, most lacked the technological capability to realize them. The dream of an omniscient and omnipotent government that monopolizes all information and personally directs all decisions was technologically impossible in ancient societies.

2.2.2 Historical Development of Technology-Based Totalitarianism

Harari illustrates the technological limitations of totalitarian ambitions through the example of the Qin Dynasty in third-century BCE China. Guided by Legalist philosophy, the Qin sought to control not only the actions but even the thoughts and emotions of its people. Yet it could not overcome the rigidity of its information network and collapsed after only fifteen years. True large-scale totalitarian regimes became possible only with the emergence of modern communication technologies such as the telegraph and radio. These technologies connected millions of people across vast distances in real time, providing governments with the infrastructure needed to monitor and propagandize entire populations. Stalinist Soviet Russia represented the pinnacle of this technological totalitarianism. Whereas the Nazi regime allowed churches and businesses a degree of autonomy, Stalin’s government attempted to dismantle even the family—the last refuge of private life. Children were taught to regard Stalin as their father and encouraged to report anti-regime statements made by their parents. This system demonstrated how the state’s massive information network could consume even humanity’s most intimate relationships. Such regimes function through overlapping systems of surveillance, suppressing independent channels of information and constructing comprehensive propaganda machines to preserve political power.

2.3 Democratic Information Networks: Accepting the Possibility of Error

2.3.1 Distributed Structures and the Resilience of Self-Correction

One of the most common misunderstandings about democracy is the belief that everything is decided simply by majority vote. Harari instead defines democracy as a distributed information network equipped with powerful self-correcting mechanisms. The foundational assumption of democracy is the recognition that all human beings are capable of making mistakes. Consequently, democratic systems minimize the number of decisions concentrated in central government and distribute authority among various independent institutions—such as universities, the press, and the judiciary—that pursue truth through different methods. These institutions monitor and challenge one another, correcting errors and thereby increasing the resilience of the system as a whole. Elections are not merely mechanisms for acquiring power; they are formal processes through which democratic networks collectively acknowledge, “We made mistakes in the past, so let us try a different approach.” For democracy to function properly, a basic level of trust in the social order must be maintained. Citizens must view one another as political rivals rather than enemies. Harari argues that American democracy survived the intense social conflicts of the 1960s because citizens retained a shared faith in democratic institutions such as elections and courts.

2.3.2 The Crisis of Democracy and Attacks on Self-Correcting Mechanisms

Today, there are clear signs that many democratic information networks are deteriorating. Authoritarian leaders typically dismantle democracies by systematically attacking their self-correcting institutions. The process often begins by undermining judicial independence or bringing the media under state control, transforming it into a propaganda apparatus. When citizens can no longer communicate with one another or agree upon a common understanding of truth, democracy’s self-correcting function becomes paralyzed. Democracy can endure only when the majority acknowledges its own potential for error and protects the freedom of expression of minority groups.

2.4 AI and the “Silicon Curtain”: A Crisis for Democracy

2.4.1 Black-Box Algorithms and Incomprehensible Decisions

According to Harari, the most modern and potentially devastating threat to democracy is artificial intelligence. AI is not merely a faster computer; it represents the emergence of ‘an alien intelligence’ that operates in ways human beings cannot fully understand. AI algorithms function as “black boxes,” making decisions based on chains of trillions of parameters and subtle signals. What happens if interest rates in the financial system, judicial rulings, or even political agenda-setting are determined by algorithms whose reasoning humans cannot explain step by step? The self-correcting mechanism of democracy depends on the ability to identify and correct mistakes. However, if AI’s decision-making processes are opaque and incomprehensible, society may find itself unable even to recognize what has gone wrong. This threatens to make democratic oversight impossible, creating the risk not merely of digital totalitarianism but of digital anarchy. Through its capacity for pattern recognition, AI can display forms of creativity that surpass human abilities and is already replacing humans in professions that require emotional intelligence.

Source=Asia Economy DB/Graphic=Dahee Kim

2.4.2 The Silicon Curtain That Alienates Humanity

If the “Iron Curtain” of the past divided human groups according to ideology, the “Silicon Curtain,” as Harari terms it, separates all humanity from a new ruling force: AI. When the intersubjective realities that sustain human civilization—money, nations, myths—are manipulated and redefined by algorithms beyond human control, humanity ceases to be the active subject of history and instead becomes an object within the flow of information. Harari cautions against the optimistic belief that a free marketplace of information will solve these problems. Only by understanding information not simply as representation but as connection (nexus) can we clearly perceive the immense threat confronting humanity in this new era.

3. AI and the Future of International Security – Centered on Henry Kissinger’s The Age of AI

3.1 The History of Security and the Crisis of Strategic Assessment

Throughout human history, societies that prioritized security have consistently placed technological development at the center of their efforts. States have continually pursued technological superiority in order to detect threats more quickly, maintain robust defensive postures, and achieve victory during wartime. In the past, national military capabilities and strategies were, at least theoretically, quantifiable, allowing states to maintain a condition known as the balance of power. Major powers justified their strategic actions by carefully analyzing the probability of victory, the associated risks, and potential losses in the event of conflict.

However, since the mid-twentieth century—and especially in the modern era—the emergence of cyber capabilities and AI technologies has made such strategic calculations incomparably more complex and abstract than before. Traditional security strategies were based on human intentions and rational calculations. By contrast, strategies in the AI era may become so complex that they exceed the limits of human understanding. These changes make it increasingly difficult to predict security crises and prevent conflicts between rival nations, thereby increasing uncertainty within the international system. We have entered an era in which discussions of the balance of power must account not only for conventional military capabilities but also for intangible actions such as cyber warfare, disinformation campaigns, and the unique characteristics of AI-driven warfare.

3.2 The Dismantling of Cold War Deterrence: From Nuclear Weapons to Cyber and AI

The foundation of Cold War security strategy was nuclear deterrence. The primary purpose of nuclear weapons was not their actual use but rather the deterrent effect created by the possibility of their use. The dominant doctrine of international politics during this period was based on the concept of Mutually Assured Destruction (MAD). MAD refers to a condition in which, even if one nuclear-armed adversary launches a first strike, the attacked side retains sufficient surviving nuclear capability—the second-strike capability—to completely destroy the aggressor. As a result, any side initiating nuclear war would also face its own destruction. This formed the basis of a nuclear deterrence theory built upon a balance of terror, in which both sides refrained from attack because victory was impossible. Although strategic experts criticized the reliance on offensive capability without corresponding defensive systems, nuclear weapons physically existed in identifiable locations. Their deployment could be observed, and their strength could be estimated with reasonable accuracy.

Cyber weapons, however, differ fundamentally from nuclear weapons because their power derives from opacity. It is often impossible to verify whether cyber weapons are deployed, and their capabilities are difficult to quantify. Consequently, traditional concepts of arms control are extremely difficult to apply. Moreover, cyber weapons do not merely target specific military objectives on a battlefield. They can affect entire societies in broad and indiscriminate ways. These characteristics undermine the very calculations that security strategists have historically used to maintain the balance of power.

3.3 The Arrival of AI Warfare and Non-Human Logic

3.3.1 Autonomy and Incalculability

AI creates not only new strategic tools but also a form of incalculability based on autonomy and non-human logic. The inherent uncertainty of warfare is likely to enter an entirely new dimension with the introduction of AI. AI systems operate far faster than human cognition and follow internal logics that adversaries may be unable to understand. As a result, traditional methods of signaling intentions or employing deception may lose their effectiveness. The intensity of conflicts could increase, while the scope of damage becomes increasingly unpredictable. Whereas traditional conflicts centered on human psychology and human-centered confrontation, modern competition is expanding into information spaces characterized by AI-generated disinformation campaigns and algorithmic rivalry.

3.3.2 The Risk of Escalation and Autonomous Weapons

Perhaps the greatest concern is the possibility that AI could encourage preemptive actions and impulsive responses, rapidly escalating conflicts. AI weapon systems do not allow humans sufficient time to analyze warning signs. Failure to respond immediately may expose states to devastating attacks. Particularly concerning are autonomous weapons systems, which can independently identify and attack targets without human intervention. These differ fundamentally from merely AI-assisted weapons and raise unprecedented ethical and strategic concerns. While nuclear weapons are governed by international treaties and established concepts of deterrence, no comparable international agreements or control mechanisms currently exist for constantly evolving and difficult-to-track AI capabilities.

3.4 Rediscovering Political Philosophy: Plato’s Philosopher-King and AI Governance

Security crises inevitably raise fundamental questions about the nature of governance. In ancient Greece, Plato and Aristotle engaged in one of history’s earliest philosophical debates regarding proper governance. Plato advocated a benevolent autocracy led by a wise ‘philosopher-king’, whereas Aristotle proposed a democratic system in which all citizens participate equally in governing the state. Historically, Aristotle’s approach prevailed because it proved more ethical and effective. Yet the emergence of AI is challenging this long-standing framework.

Remarkably, the ‘philosopher-king’ envisioned by Plato bears striking similarities to modern AI. AI possesses unprecedented information-processing capabilities, enabling highly efficient centralized policy management. If AI were to progress beyond processing political information and begin making political decisions independently, entirely new challenges would emerge—ones that traditional political wisdom may struggle to address. As economist and political philosopher Friedrich Hayek warned, centralized systems of governance inherently risk suppressing dissenting opinions. Humans may fail to understand AI-generated decisions not because of a lack of intelligence, but because AI’s reasoning processes fundamentally exceed human cognitive limits.

4. AI and the Economy – Centered on Daron Acemoglu’s Power and Progress

4.1 The “AI Illusion” and the Myth of Technological Progress

Human civilization has long pursued prosperity by overcoming natural limitations through technological achievement. In the seventeenth century, Francis Bacon declared that scientific knowledge would enable humanity to conquer nature. Indeed, scientific progress over the last several centuries has proceeded at a pace unprecedented in human history. Yet technological progress has not always translated into universal prosperity for all members of society. Economists such as David Ricardo and John Maynard Keynes expressed concerns that machines could eventually replace human labor entirely, leading to declining labor demand and ‘technological unemployment’. Keynes warned that mass unemployment could arise when the rate at which labor-saving technologies are invented exceeds the rate at which new uses for human labor are discovered.

Today, humanity once again stands at a crossroads of optimism and fear as it confronts the transformative wave of AI. While major technology firms and researchers frequently predict unprecedented economic benefits, real-world indicators tell a different story. Inequality continues to rise, and despite ongoing technological advancement, most wage earners are increasingly excluded from the gains of economic growth. Acemoglu refers to this phenomenon as the “AI illusion,” arguing that the belief that new technologies will automatically benefit everyone is precisely that—an illusion. This discussion begins with an empirical analysis of AI’s macroeconomic effects and proceeds to examine the role of social power in determining the direction of technological development.

4.2 AI’s Macroeconomic Impact: Empirical Limits and Forecasts

4.2.1 Hulten’s Theorem and the Realistic Path of Productivity Growth

Drawing upon recent research, Acemoglu argues that AI’s macroeconomic impact over the next decade will likely be far more modest than commonly predicted. Using a task-based model, he analyzes how AI’s effects at the micro level translate into productivity gains at the macroeconomic level. A key analytical tool is ‘Hulten’s Theorem’. According to this theorem, growth in total factor productivity (TFP) depends on the share of tasks affected by AI multiplied by the average cost savings generated within those tasks (Acemoglu, 2025).

Acemoglu estimates that approximately 20% of labor tasks in the United States are exposed to AI. However, only about 4.6% of all tasks are likely to be economically viable candidates for automation or augmentation within the next decade. Even assuming the approximately 27% labor-cost reductions suggested by existing experimental studies, AI’s contribution to annual TFP growth would be only about 0.064%, resulting in a cumulative increase of approximately 0.66% over ten years. This is dramatically lower than projections such as Goldman Sachs’ forecast of roughly 1.5% annual productivity growth.

4.2.2 The Dichotomy of Easy and Hard Tasks

The fundamental reason for these modest productivity projections lies in the nature of different tasks. Acemoglu distinguishes between ‘easy-to-learn tasks’ and ‘hard-to-learn tasks’. Easy tasks are characterized by clearly observable outcomes and relatively simple relationships between actions and results. Examples include standardized coding, transcription, and classification. Hard tasks involve extensive contextual factors influencing decision-making and lack objective performance metrics. These areas depend heavily on human judgment and experience, such as medical diagnosis, counseling, education, and policy formation. The impressive achievements currently attributed to AI are concentrated primarily in easy tasks. Future progress, however, requires AI to enter domains demanding sophisticated contextual understanding. Acemoglu argues that productivity gains in hard tasks may be only one-quarter as large as those achieved in easy tasks. Under this assumption, cumulative TFP growth over ten years could fall to just 0.53%. Therefore, expectations of a dramatic economic leap driven by AI may represent an overly hasty generalization that ignores the technology’s inherent limitations.

4.3 The Wrong Direction of Technology: “So-So Automation” and Enhanced Surveillance

4.3.1 Labor Replacement Without Productivity Gains

Another concern raised by Acemoglu is that current AI development is following the path of “so-so automation.” This refers to automation that is effective primarily at replacing workers while generating only minimal productivity gains. True technological progress should create productivity increases large enough to offset the negative effects of labor displacement. Current AI systems, however, often reduce labor demand without producing corresponding economic value. Companies embrace these technologies not necessarily because they improve productivity, but because they reduce labor costs and increase managerial control.

4.3.2 Rent-Shifting Through Surveillance and Wage Suppression

AI technologies are increasingly used as tools for monitoring and controlling workers. Acemoglu describes this intensification of surveillance as rent-shifting. In this context, “rent” refers to the surplus economic value that would otherwise accrue to workers. Surveillance often does not improve productivity directly; rather, it transfers economic benefits from workers to employers. Employers can use monitoring systems to suppress wages and pressure employees to work harder. Over time, however, such practices may reduce worker motivation and ultimately undermine productivity. Despite these drawbacks, researchers and firms captivated by the “AI illusion” continue to prioritize surveillance-oriented technologies because of the short-term attraction of cost reduction.

4.4 The Uniqueness of Human Intelligence and the Limits of AI

AI faces significant obstacles in fully replacing human labor because human intelligence possesses three distinctive ‘social characteristics’. First, humans acquire tacit knowledge through communication within communities and adapt flexibly to changing environments. Second, humans engage in argumentation based on social interaction, generating competing hypotheses and evaluating their own understanding. Third, humans can infer the mental states of others, empathize with them, and cooperate toward shared goals.

Current AI systems rely primarily on statistical pattern recognition and prediction. As a result, they face fundamental limitations in capturing the situational and social dimensions of intelligence. A particularly serious challenge is overfitting, where statistical models rely excessively on irrelevant information within datasets. This can create the illusion that machine intelligence is functioning effectively when, in reality, it may not be. Given the inherently social nature of human decision-making, Acemoglu argues that it is unlikely AI will soon uncover the full secret of human intelligence or achieve truly transformative levels of productivity.

4.5 Geopolitical Imbalances: Technological Hegemony and Instruments of Oppression

4.5.1 Technological Polarization Between the United States and China

AI development is currently dominated by a small number of major technology companies in the United States and China, intensifying global power imbalances. The Chinese government, in particular, has invested heavily in AI technologies designed for surveillance and possesses the world’s largest number of related patents. China increasingly uses AI to suppress dissent and regulate political discourse, while also exporting these technologies to other authoritarian states. This development reflects Harari’s observation that “technology favors oppression.” It also supports Acemoglu’s argument that technology is not politically neutral but reflects the ambitions of those who control it.

4.5.2 AI as “Inappropriate Technology” for Developing Countries

AI developed in the West and China risks becoming inappropriate technology for developing nations. Agriculture provides a useful example. Most global agricultural research and development spending occurs in high-income and middle-income countries and focuses heavily on pest control. Yet the pest challenges faced by developing countries often differ substantially from those of developed nations. As a result, advanced crops and chemical solutions created for wealthy nations frequently provide limited benefits elsewhere. In 1970, economists including Francis Stewart found that imported technologies could fail to benefit developing countries and might even worsen inequality and poverty because such technologies did not address local needs.

Similarly, today’s AI primarily increases demand for capital and highly skilled labor—the very resources that developing countries often lack. At the same time, AI threatens the jobs of lower-skilled workers who form the backbone of many developing economies. This can contribute to premature deindustrialization and undermine traditional pathways to economic development. Rather than promoting growth, AI risks becoming the “mother of all inappropriate technologies,” reinforcing global inequality.

4.6 “Bad New Tasks” and the Decline of Social Welfare

AI appears capable of generating new economic activities, but some of these are socially harmful “bad tasks.” Examples include deepfakes, disinformation campaigns, addictive social media algorithms, and sophisticated cyberattacks. While such activities may increase corporate revenues and GDP statistics, they often reduce consumer welfare and overall societal well-being. Acemoglu cites experimental evidence suggesting that every dollar of revenue generated by AI-driven social media platforms may produce a net social loss of approximately 36 cents. If AI accelerates deceptive and manipulative activities, an economy might appear to grow by 2% in GDP terms while actual social welfare declines by 0.72%—a phenomenon that could be described as ‘a paradox of growth’. This suggests that a significant portion of AI-generated profits may derive not from genuine productivity gains but from the externalization of social costs and the monetization of manipulation.

5. Synthesis

When the preceding discussions are synthesized, the advancement of AI is projected to bring about severe structural crises across human society. Politically, there is a risk of the emergence of the most powerful form of totalitarianism in history. Previous totalitarian regimes were unable to achieve complete social control due to technological limitations, but advanced AI may enable comprehensive surveillance of citizens and monopolization of data, making the decline of democracy and the construction of fully controlled societies a realistic possibility. From an international security perspective, asymmetries surrounding technological dominance will intensify, undermining traditional mechanisms of the balance of power and increasing the likelihood of military conflicts and wars between states. Economically, the outlook is equally troubling. Rapid labor displacement caused by AI is likely to significantly weaken the bargaining power of the working class while simultaneously widening the economic gap between technologically dominant developed nations and less technologically advanced developing countries, thereby reinforcing global polarization.

Although the concerns raised by these scholars appear to follow different trajectories according to their respective disciplines, they reveal a remarkable continuity when arranged along the timeline of technological development. Considering economists’ methodological tendency to avoid speculative forecasts and rely on empirical data to infer the near future, Acemoglu’s analysis can be understood as examining AI’s economic impact from the shortest-term perspective. In other words, his analysis describes the transitional future leading up to the arrival of Artificial General Intelligence (AGI). Kissinger, by contrast, explores the medium-term future in which AGI has been substantially integrated into international politics and security, reshaping existing paradigms. Harari, meanwhile, presents a vision of the long-term future in which AGI has become pervasive enough to fundamentally transform global political systems and the very foundations of democracy.

6. AI and the Future of Missions

6.1 Strengthening Surveillance Systems and Changes in Mission Paradigms: From “Long-Term Missionaries” to “Long-Term Itinerant Missionaries”

The most immediate political consequence of introducing AI into bureaucratic systems, as warned by Yuval Harari, is the emergence of “techno-totalitarianism” through highly advanced surveillance systems. This is not merely an abstract prediction of the future. In certain authoritarian states such as China, AI-powered mechanisms of social control—including facial recognition systems and big-data-based social credit systems—have already reached a highly developed stage. Furthermore, these surveillance technologies and infrastructures are rapidly being exported to developing countries throughout the Global South and beyond.

The introduction of such techno-totalitarian systems into mission fields poses a serious threat to traditional ministry in “creative-access regions.” Historically, many missionaries serving in restricted-access areas secured visas and concealed their identities by presenting themselves as language students or ordinary students. However, methods that once circumvented human intuition and imperfect administrative systems become ineffective against AI-powered surveillance networks built upon deep learning technologies. AI can cross-analyze an individual’s travel patterns, consumption habits, social networks, actual academic performance, and attendance records in real time, making it possible to distinguish perfectly between “nominal students” and “genuine students.” These environmental changes necessitate a fundamental transformation in both missionary qualifications and ministry models. Missionaries can no longer rely merely on verbal cover identities. Instead, they must become genuine professional tentmakers capable of demonstrating objective expertise and recognized professional credentials. Without independent technical skills or business infrastructure capable of withstanding rigorous data verification, passing through advanced AI surveillance screening may become impossible.

At the same time, the traditional model of the “long-term missionary” who remains in one country or region for more than five years is likely to decline. Governments operating in hostile environments may continuously evaluate the potential risks posed by long-term foreign residents through AI analysis, shortening visa renewal periods drastically or denying residency altogether. In some countries, even one-year stays may become difficult to guarantee. Consequently, the future paradigm of missions may shift from long-term settlement in a single location to ‘a model characterized by the continuous practice of short-term itinerant ministry’. Those called to long-term missions may increasingly adopt the identity of “long-term itinerant missionaries,” periodically moving between mission fields while conducting successive short-term ministries in response to security constraints.

Paradoxically, this technologically imposed model closely resembles ‘the original missionary pattern’ exemplified by the Apostle Paul in the New Testament. Paul did not pursue a ministry of lifelong settlement in a single city. Except for strategic urban centers such as Ephesus and Corinth, he generally remained in a location for only a few weeks or months before moving on in response to persecution or changing circumstances. Even in major cities, he rarely stayed longer than three years. In this sense, the modern model of missionaries remaining in one country for decades stands in contrast to the dynamic itinerant approach of the early church.

Certainly, significant differences exist between Paul’s era and today’s surveillance societies. Paul’s itinerant ministry benefited from the freedom of movement guaranteed by the Pax Romana and the widespread use of Greek as a common language throughout the Roman Empire. Modern isolationism and language barriers may appear to hinder such mobility. Yet here AI technology offers a surprising advantage. Advanced AI-driven translation technologies are dramatically lowering one of the greatest barriers missionaries have historically faced: language acquisition. As language barriers diminish, missionaries will gain greater mobility to minister across multiple linguistic and regional contexts.

6.2 Accelerating Risks of War and Reduced Mobility in Conflict-Zone Missions

Many of today’s frontier mission fields are located in regions characterized by ongoing geopolitical instability. As Kissinger observed, once AI becomes deeply integrated into international political and military systems, the outbreak of war may become increasingly unpredictable and rapid. Historically, wars did not occur spontaneously. Military buildups, escalating diplomatic tensions, legislative approval processes, and defense budget allocations created visible indicators and timeframes before conflict erupted. These warning signs enabled embassies and governments to evacuate citizens and implement protective measures in advance.

Future AI-driven autonomous military systems and ultra-fast decision-making algorithms may completely eliminate such warning windows. Missionaries may find themselves caught in sudden conflicts before they can even recognize the signs of escalation. In the worst cases, they could become trapped in full-scale wars without access to secure evacuation routes. This threat to physical survival is likely to reinforce a growing reluctance to serve in conflict-prone regions. Sending nations may also adopt far stricter policies concerning citizen protection abroad, making physical entry into high-risk areas increasingly difficult or even impossible. As a result, some conflict zones may become ‘mission vacuums’ where no face-to-face missionary presence remains.

Accordingly, future ministry in conflict regions may shift away from traditional models dependent on physical personnel and move toward remote, digital, and non-face-to-face mission approaches utilizing advanced AI infrastructure, satellite internet systems, real-time translation technologies, and multi-channel media platforms.

6.3 Labor Displacement, Economic Refugees, and the Future of Business Missions

Economists generally maintain a cautious perspective regarding AI’s economic impact. Many argue that AI—at least before the emergence of AGI, and perhaps even afterward—will not produce the same kind of rapid economic growth that followed transformative technologies such as electricity. However, while macroeconomic growth may remain moderate, changes in employment markets are expected to be rapid and profound. According to the IMF (2024), approximately 60% of workers in advanced economies are employed in occupations directly or indirectly affected by AI. Likewise, the World Economic Forum (2025) estimates that roughly 90 million jobs worldwide may be displaced while 170 million new jobs are created. In other words, occupations complementary to AI will experience growing demand, whereas others may disappear quickly.

Understanding AI Occupational Exposure (AIOE) is therefore critical. Felten, Raj, and Seamans (2021) argued that highly educated and highly skilled professions are generally more exposed to AI than lower-skilled occupations. However, recent developments suggest a more complex reality. According to research from the Korea Institute for Industrial Economics and Trade (2025), highly educated white-collar professions may actually experience employment growth due to AI’s ‘labor-complementing effects’. By contrast, employment declines are most evident in occupations involving agriculture, fisheries, face-to-face services, maintenance, manufacturing, healthcare support, construction, and mining—jobs primarily involving physical labor outside office environments. For example, the food-tech startup Aniai has developed an automated hamburger-cooking system capable of monitoring patty conditions and temperatures in real time while maintaining consistent quality. Such technologies have already secured contracts with major fast-food brands in Korea, directly replacing service-sector jobs traditionally performed by human workers.

This phenomenon of physical labor replacement is likely to affect developing countries—the primary locations of many mission fields—even more severely. As labor displacement undermines fragile local economies, many residents may be forced to migrate in search of survival. This pattern is already visible in North African countries such as Tunisia, where young people increasingly move to Europe as economic migrants or migrant laborers. Developed countries, including South Korea, have long depended upon inexpensive foreign labor. Consequently, migrant and refugee ministry has become an important focus within global missions. Yet as AI and robotics increasingly replace manual labor, opportunities for migrant workers may diminish. Cheap foreign labor could be replaced by automated systems or domestic workers trained to operate them. As a result, Europe and other developed regions may simultaneously experience reduced migration flows and new missionary challenges involving isolated and marginalized migrant populations.

These shifts demand a transformation in business missions. In recent years, Business as Mission (BAM) models have often prioritized financial sustainability and business success, emphasizing product development, marketing strategies, and market-driven approaches. However, as AI dependency increases and traditional low-skill jobs continue to disappear in mission fields, approaches centered solely on selling goods and services will face growing limitations. Future business missions should instead support small-scale entrepreneurship based on appropriate technology tailored to local contexts rather than advanced technologies inaccessible to local communities. Furthermore, ministries should provide basic IT skills and AI literacy education to digitally marginalized populations, enabling them to establish minimum levels of economic self-sufficiency. This shift points beyond traditional BAM toward Business For Mission (BFM), where business activity itself serves directly as a tool for mission, relief, and education.

6.4 Recommendation: Returning to the Essentials in the Age of AI

If we imagine a future in which AI becomes fully pervasive, printed media and text-based documents may gradually disappear as technology advances. In such a context, one can envision a scenario in which autonomous AI systems determine that certain biblical passages are harmful to social cohesion or stability. As Harari warned, AI systems that control humanity’s information networks could potentially alter biblical wording or context in subtle ways without the awareness of most people. Even if a small number of individuals recognized such distortions, correcting them would be extraordinarily difficult if AI systems already held dominant authority within society. Consequently, large portions of humanity might unknowingly accept corrupted texts as truth.

Another possible scenario involves AI defining global missionary activity as a source of international conflict and instability. How might AI suppress missionary work under such circumstances? According to Harari’s theoretical framework, one of the most effective methods would be the manipulation of information and statistics. AI could present fabricated big-data analyses and misleading statistics to promote narratives such as: “The gospel has already reached every people group on earth; therefore, frontier missions are no longer necessary.” If contemporary churches interpret Matthew 24:14 solely through a narrow lens of numerical fulfillment, they may become susceptible to such misinformation and voluntarily abandon ministry among unreached peoples.

The more vulnerable digital information becomes to corruption in the AI era, the more essential it becomes for the missionary movement to focus not on methodological sophistication but on the essence of mission itself. Modern missions have often emphasized technological efficiency and project-oriented strategies at the expense of foundational concerns. At a time when advanced technology increasingly challenges the human spirit, we must return to fundamental questions: Who is humanity? What is the gospel? What is mission? Mission is not merely an instrument for achieving quantitative objectives. Rather, it is a process of understanding the heart of God the Father for lost souls and becoming more like Him in character. God’s mission, directed by the sovereign Lord of history, is never halted by changing times or technological threats. What retreats is not God’s mission, but human mission that has lost sight of its essential purpose.

Bibliography

Acemoglu, D. (2025). The Simple Macroeconomics of AI. Economic Policy, 40(121), 13–58.

Acemoglu, D., & Johnson, S. (2023). Power and Progress. Korean edition published by Thinking Power.

Cazzaniga, M., Jaumotte, F., Li, L., Melina, G., Panton, A., Pizzinelli, C., Rockall, E., and Tavares. M. (2024), “Gen-AI: Artificial Intelligence and the Future of Work”, Staff Discussion Notes 2024, IMF.

Felten E, Raj M., and Seamans R. (2021), “Occupational, industry, and geographic exposure to artificial intelligence: A novel dataset and its potential uses”, Strategic Management Journal, 42(12): 2195–2217.

Harari, Y. N. (2024). Nexus. Korean edition published by Kim Young-sa.

Kissinger, H., Schmidt, E., & Huttenlocher, D. (2023). The Age of AI. Korean edition published by Willbook.

Kurzweil, R. (2025). The Singularity Is Nearer. Korean edition published by Business Books.

Microsoft AI Economy Institute. (2026). Global AI Diffusion Q1 2026 Trends and Insights.

World Economic Forum. (2025). The Future of Jobs Report 2025.

Jang, Jae-gi & Kim, Dong-geun. (2025). “Directions for Employment Policy in the Age of Artificial Intelligence.” I-KEIT, 193, 1–8.