OVERVIEW OF THE AUTHORS’ WORK

Nick Bostrom, a Swedish philosopher at the University of Oxford, is widely recognized for his contributions to the study of existential risk, the anthropic principle, the ethics of transhumanism, and the challenges posed by superintelligence. For his part, Milan M. Ćirković, a Serbian astronomer and astrophysicist, focuses his research on astrobiology, global catastrophic risks, and the future of humanity—a field in which he has collaborated closely with Bostrom.

EXCERPT 001. INTRODUCTION TO THE BOOK “GLOBAL CATASTROPHIC RISKS”

Nick Bostrom and Milan Ćirković conduct a study on global catastrophic risks, defined as events capable of causing severe harm to human well-being on a planetary scale. Based on this concern, the authors examine what should be understood by “risk” to humanity and propose a classification of these threats, distinguishing between natural risks, risks arising from unintended consequences, and risks caused by hostile acts.

Although the concept of global catastrophic risk does not have a single, strict definition, it generally refers to events capable of causing damage of enormous magnitude to the human species across the entire planet. These hazards include “volcanic eruptions, pandemic infections, nuclear accidents, global tyrannies, scientific experiments, climate change, cosmic risks, and economic collapse” (Bostrom & Ćirković, 2008).

Based on this conception, the authors argue that a comprehensive approach to global catastrophic risks is both coherent and advantageous, since, despite the diversity of existing threats, many of them share common causal mechanisms, effects, and problems. Consequently, they should not be studied in isolation but rather as part of a single field of research.

This interrelationship can be illustrated through various examples. First, a significant portion of the damage caused by destructive phenomena stems from their second-order social effects. Thus, risks such as nuclear terrorism or pandemic diseases not only cause direct losses but can also trigger processes of social disintegration and institutional collapse. Second, events such as large asteroid impacts, supervolcanic eruptions, or nuclear war share the ability to eject enormous amounts of soot and aerosols into the atmosphere, significantly altering the global climate (Bostrom & Ćirković, 2008). These connections justify the need to analyze multiple types of risks jointly. Similarly, the authors highlight the importance of taking an integrated approach to the methodological, conceptual, and cultural issues associated with their study.

There are also pragmatic reasons for considering global catastrophic risks as a single field of research. Because these risks typically receive little attention and entail high management costs, studying each threat in isolation can lead to an inefficient allocation of resources, focusing the attention of experts and the public on certain risks while others—potentially more severe or more difficult to mitigate—remain neglected.

As a result, partial or inadequate prevention policies may be implemented that, far from reducing the overall risk, actually foster the emergence of new vulnerabilities and increase the general level of threat. From this perspective, a broad view is essential for setting priorities and designing effective mitigation strategies.

Based on these arguments, the book’s objective is to examine the range of global catastrophic risks that humanity faces today or may face in the future. This work is intended for readers with an interdisciplinary background, and its purpose is to foster research, public awareness, and political debate regarding the management of these risks. In the view of Bostrom and Ćirković, the development of an interdisciplinary community dedicated to the study of global catastrophic risks will increase the likelihood of designing and implementing appropriate responses to the major challenges of the 21st century.

Taxonomy and Organization

Once the general concept of global catastrophic risk has been established, the authors specify which types of events can be considered within this category. The fundamental criterion is that the resulting damage must reach a truly global magnitude. In this sense, a catastrophe that caused 10,000 deaths or economic losses of 10 billion would not constitute a global catastrophic risk (Bostrom & Ćirković, 2008). In contrast, an event responsible for approximately 10 million deaths or losses in the order of 10 trillion could indeed be classified as such (Bostrom & Ćirković, 2008).

The authors also point out that history offers numerous examples of large-scale catastrophes. Among these are the An Lushan Rebellion, the Taiping Civil War, the famine caused by the Great Leap Forward in China, the Black Death in Europe, the Spanish flu pandemic, World War I and World War II, the Nazi genocide, the great famines in India, Stalin’s totalitarian regime, and the drastic reduction in the Native American population as a result of smallpox and other diseases introduced by European colonizers, among other cases.

To assess the severity of a risk, Bostrom and Ćirković propose considering three fundamental dimensions. The first is scope, that is, the number of people and species potentially affected. The second is intensity, understood as the magnitude of the damage caused. Finally, the third dimension is probability, which refers to the likelihood that the event will occur based on the available evidence.

Based on these variables, the authors develop a more detailed classification. Depending on its scope, a risk can be personal, when it affects an individual; local or global, when it affects entire populations; and transgenerational, when its consequences extend to future generations. In terms of intensity, they distinguish between imperceptible risks; bearable risks, which cause significant damage but are compatible with the continuity of civilization; and terminal risks, whose effects entail death or a permanent and drastic reduction in quality of life. Within this classification, global catastrophic risks correspond primarily to threats of global or transgenerational scope and of bearable or terminal intensity.

As shown in Figure 1, this taxonomy can be expanded to incorporate conceptually possible risks that fall outside conventional categories. In the case of transgenerational risks, the authors consider the possibility of threats so severe that they would not only compromise the well-being of future generations but also the future of the universe’s “cone of light”—understood as the region of spacetime from which new forms of intelligent and conscious life could emerge. Furthermore, some theories of value hold that certain states of existence could be worse than death. Examples of this would include permanent and extreme forms of slavery or mind control. From this perspective, the axis corresponding to risk intensity could extend even beyond the terminal category, incorporating scenarios of irreversible suffering that transcend mere physical extinction (Bostrom & Ćirković, 2008).

A subset of global catastrophic risks consists of existential risks, understood as those capable of causing the extinction of intelligent life originating on Earth or of permanently and drastically reducing its quality of life. Unlike other types of threats, existential risks are characterized by their irreversible nature: once they materialize, there is no possibility of recovering from their consequences or learning from the experience to prevent future disasters. Therefore, the only rational strategy is to prevent them from occurring in the first place.

The severity of an existential catastrophe depends, to a large extent, on the assumptions adopted by value theory, since it is necessary to determine what ethical weight should be given to future generations. At the same time, the assessment of these risks presents particular methodological difficulties related to observational selection effects and the need to avoid so-called anthropic bias—a concept Bostrom uses to refer to the uncertainty surrounding our position within the universe.

In this context, the authors argue that it is essential to develop a systematic framework for the study of major risks, explicitly incorporating the temporal dimension. Understanding the evolution of these threats allows us to trace the origins of contemporary challenges and assess their potential future consequences. Thus, for example, the analysis of nuclear terrorism should not be limited to calculating the probability of an accident or an attack, but must also consider how such a risk could transform the future course of civilization. Similarly, climate change resulting from greenhouse gas emissions does not, in and of itself, constitute an immediate catastrophic risk; the real problem lies in the cumulative effects that these emissions may generate in the long term. Similarly, anticipating the risks associated with technological development is a central element of the analysis.

The authors also point out that, under certain circumstances, it may be useful to examine scenarios that even appear physically impossible. An example of this is the hypothetical risk associated with experiments conducted in particle colliders (Bostrom & Ćirković, 2008). From a strictly objective perspective, it is likely that such experiments cannot trigger a global catastrophe, so their risk would be practically zero. However, the absence of absolute certainty prevents us from completely ruling out that possibility. Consequently, as long as a reasonable margin of uncertainty remains, the risk must be considered subjective, as it reflects the limits of available knowledge. Because this uncertainty is relevant to the analysis, the authors incorporate both objective and subjective risks into their definition of global catastrophic risks.

Nevertheless, distinguishing between objective and subjective risk is not always straightforward. An illustrative example is the impact of an asteroid on Earth. If there were complete certainty that no asteroid would collide with the planet over a sufficiently long period, it could be asserted that the objective risk is nonexistent. However, such certainty would only be possible after precisely calculating the trajectories of all potentially hazardous objects. As long as that information remains incomplete, the risk continues to exist from a subjective standpoint, simply reflecting our state of ignorance (Bostrom & Ćirković, 2008). The authors illustrate this difference using the example of an apparently empty cave. Although it objectively poses no danger, a person who believes a lion lives inside will act rationally by avoiding entering it, since, given their beliefs, the expected harm outweighs the potential benefit.

In the case of asteroid impacts, the risk can be estimated using empirical data. It is possible to calculate the probability of a catastrophic collision using historical impact records and astronomical models that describe the behavior of near-Earth objects. Consequently, the assessment of such threats requires rigorous scientific research, and the probabilities obtained depend directly on the available evidence. However, there are other risks for which neither empirical data nor scientific models allow for reliable probabilistic estimates.

This is the case, for example, with the possibility of a terrorist attack using biological weapons over the next decade or with the eventual emergence of a totalitarian regime with global reach before the end of the 21st century. In these cases, it is impossible to establish probabilities using strictly scientific methods. Therefore, the analysis relies heavily on arguments of plausibility, historical analogies, and expert judgments (Bostrom & Ćirković, 2008).

Although the authors acknowledge the importance of using rigorous methods whenever possible, they caution that limiting the analysis solely to those risks amenable to precise quantification would lead to an overly restrictive application of the scientific method. As a result, many threats that could be decisive for the future of humanity would be excluded from the debate. Furthermore, this stance would create a false dichotomy between “scientific” risks and “speculative” risks, when in reality both are part of an analytical continuum.

Based on these considerations, Bostrom and Ćirković identify a systematic bias toward the study of relatively minor threats—ones that can be analyzed using well-established scientific tools—while potentially more serious risks tend to be sidelined due to their speculative nature. To correct this imbalance, they structure the book into four main sections: a first section devoted to conceptual background; a second focused on natural risks; a third centered on risks arising from unintended consequences; and a fourth dedicated to the study of risks caused by hostile acts.

This classification, however, should not be interpreted rigidly. An earthquake, for example, is generally considered a natural phenomenon because it results from the movement of tectonic plates, a process beyond human control. Nevertheless, the level of associated risk also depends on social decisions, such as the location of cities, building codes, and the resilience of infrastructure. If human settlements were located far from geological faults and consisted of earthquake-resistant buildings, the consequences of an earthquake would be considerably less severe. In this sense, risk arises from the interaction between natural processes and human decisions.

A similar situation applies to nuclear weapons. Although this is a technology created by humans, the effects of an explosion depend largely on natural factors such as wind, temperature, and precipitation, which determine the dispersion of radioactive fallout and the likelihood of firestorms. The ultimate magnitude of the disaster is, therefore, the result of the interaction between these two types of factors.

Finally, the authors emphasize that the nature of a risk can change over time. The case of famines is illustrative: historically, they were primarily associated with natural phenomena, such as climate variations or changes in food availability. Today, however, most major famines stem from market failures, armed conflicts, political crises, or processes of social collapse. Consequently, the same phenomenon can shift from the category of natural risk to that of an unintended consequence or hostile act, depending on the historical context and the social conditions in which it occurs.

Part I. Background

The purpose of this first part of the book is to provide a general context and methodological guidance for thinking systematically and critically about global catastrophic risks. The opening chapter, written by American astrophysicist Fred Adams, presents a long-term scenario for the future of Earth, the galaxy, and the universe as a whole.

It is estimated that, in approximately 3,500 million years, the Sun’s increased luminosity will have sterilized Earth’s biosphere, although complex life will likely disappear much sooner (Bostrom & Ćirković, 2008). This would be the foreseeable fate of life on the planet if no other events were to intervene. However, it is reasonable to assume that, if humanity and its technological civilization were to survive for a sufficiently long period, by that time they would have developed the capacity to colonize space.

If a catastrophic event were to cause the extinction of Homo sapiens and a large portion of terrestrial organisms, there would still be a window of approximately one billion years for another intelligent species to evolve and take humanity’s place. However, it is unknown whether a new evolutionary trajectory would produce a species comparable to humans—one that is self-aware and capable of developing a technological civilization.

If intelligent life were to expand into space through the use of technology, its existence could be prolonged for an extraordinarily long period (Bostrom & Ćirković, 2008). Even so, at some point the universe would have to come to an end. The future conceived on these time scales far exceeds the ordinary capacity of human understanding.

After examining this distant fate, the authors turn their gaze to the deep past. Some of the great cataclysms that have occurred throughout planetary history have left identifiable geological records. Among these, the Permian–Triassic mass extinction stands out, considered one of the most severe biological crises in Earth’s history.

Phenomena such as asteroid and comet impacts, as well as large-scale volcanic eruptions, have contributed to various mass extinctions. Other processes, such as variations in the intensity of solar radiation, can also exert extreme pressures on ecosystems. According to the authors, all mass extinctions appear to have been mediated by atmospheric changes, such as shifts in air composition or global temperature (Bostrom & Ćirković, 2008).

However, not all species have disappeared as a direct result of natural cataclysms. Some were displaced due to competition with other species for available resources and ecological niches. This observation helps us understand that extinction can result from both abrupt events and gradual evolutionary processes.

Chapter 3 specifically examines some of the mechanisms of evolutionary transformation. It is believed that modern humans coexisted for a certain period with other hominid species, including Neanderthals, who manufactured and used composite tools of remarkable complexity. It also mentions Homo floresiensis, a species discovered on an Indonesian island and popularly known as “the hobbit” due to its short stature. These cases demonstrate that the extinction of species with advanced cognitive abilities has already occurred in the history of the planet.

In Chapter 4, James Hughes analyzes the specific cognitive tendencies that arise in the face of expectations of apocalyptic disasters. The author examines what he terms apocalyptic, psychocultural, or utopian groups of a millenarian nature, as well as the dynamics of eschatological beliefs and behaviors associated with them (Bostrom & Ćirković, 2008). Studies on millenarianism reveal that these tendencies appear in very different cultures and are not a phenomenon exclusive to any particular society.

Hughes shows that these ideas can be studied through various forms and tropes found in Europe, India, and China. According to the author, it is possible to conduct rational and technocratic analyses of the probabilities of achieving a future free from disease, hunger, or death, just as it is possible to assess threats such as wars, pandemics, or asteroid impacts. However, even seemingly probabilistic studies are often permeated by millenarian tendencies, whether positive or negative, fatalistic or messianic.

Although eschatological rhetorical devices may respond to legitimate social needs and help mobilize necessary actions, they can also foster processes of social disengagement. For this reason, Hughes argues that vigilant and historically informed self-examination is required to keep efforts focused on developing concrete responses to real challenges (Bostrom & Ćirković, 2008). This caution is especially important because thinking and acting rationally in the face of global and existential catastrophic risks is often particularly difficult.

Along these same lines, Eliezer Yudkowsky, a researcher specializing in artificial intelligence, points out that extremely high mortality rates, as well as qualitatively different scenarios—such as the complete extinction of humanity—trigger a different mode of reasoning, as if one were entering a separate mental realm (Bostrom & Ćirković, 2008). Perception of the problem shifts even further when one recognizes that some existential risks may be caused by humanity itself, as this possibility can lead to the belief that the human species does not deserve to survive.

Psychologists and economists have developed an extensive body of empirical literature on the heuristics and biases present in human cognition. Yudkowsky applies these findings to the analysis of large-scale risks. Among the phenomena he examines are the availability heuristic, hindsight bias, so-called black swans, the conjunction fallacy, confirmation bias, anchoring, adjustment and contamination, the affect heuristic, insensitivity to scale, calibration problems, overconfidence, and bystander apathy (Bostrom & Ćirković, 2008).

Along with these general cognitive biases, anthropic bias must also be considered. This bias differs from those studied by Yudkowsky in that it has a more theoretical foundation and applies specifically to certain classes of inferences. It arises when relevant observational selection effects are not taken into account. Such effects stem from the fact that the available evidence is conditioned by the existence and position of the observer interpreting it. Ignoring this circumstance can lead to errors in the probabilistic evaluation of important hypotheses.

In Chapter 6, Milan Ćirković analyzes the applications of observational selection theory to the study of catastrophic and existential risks, even though some of these applications may not be obvious at first glance. One example involves inferring that certain existential disasters are extremely improbable simply because there are no observable precedents.

This line of reasoning can be misleading, since the only places from which an observation can be made are precisely those that have not been completely destroyed and where intelligent observers still exist. Consequently, the fact that Earth has not yet suffered an observable existential catastrophe does not prove that such events are inherently improbable. Even if the destruction of planets or intelligent species were relatively frequent, observers could only be found in those places where such destruction had not yet occurred.

Other applications of the anthropic principle, such as the so-called Doomsday Argument—originally formulated by Brandon Carter and later developed by John A. Leslie—are of more questionable validity, especially in their generalized forms. This argument aims to make a probabilistic prediction about the total number of members the human species will have based on the temporal position each individual occupies within it. Ćirković believes that this reasoning must be applied with significant restrictions.

Beyond these theoretical controversies, there are professional communities that assess risks on a daily basis. The next two chapters of the book present, respectively, the perspectives of systems engineering and the insurance industry (Bostrom & Ćirković, 2008).

In Chapter 7, Yacov Y. Haimes proposes various strategies for organizing the analysis of risk variables in projects involving complex systems. The author asks what kind of knowledge is required to make appropriate decisions regarding risk mitigation. According to Haimes, understanding the complex and partially unknowable nature of emergent systems requires modelers and thinkers who are free from bias, capable of experimenting with multiple modeling and simulation approaches, and able to collaborate in the search for appropriate solutions (Bostrom & Ćirković, 2008).

Haimes warns that organizing the analysis solely based on the expected value of risk may prove insufficient. This measure combines the probability of an event with the magnitude of its consequences, but it can obscure extreme scenarios with low probability and high impact. Therefore, decision-makers require a more precise breakdown that allows them to analyze separately the probability of consequences with different levels of severity. To this end, Haimes proposes his multi-objective risk method.

In Chapter 8, Peter Taylor analyzes the relationship between the insurance industry and global catastrophic risks. Insurance companies enable individuals and organizations to manage the financial consequences of certain hazards through risk transfer and distribution mechanisms.

However, the ability to privately insure against a catastrophe is limited by both the magnitude and the nature of the risk. Although the insurance industry has devoted relatively little attention to global catastrophic risks, it has extensive experience in dealing with smaller-scale threats. Some of its concepts and methods can therefore be adapted to the analysis of larger-scale risks.

Taylor particularly emphasizes the importance of uncertainty. A stochastic model can represent certain events using a probability distribution that accounts for various possible outcomes. However, in addition to the randomness inherent in the model itself, two other sources of uncertainty must be considered: that related to the values of the parameters entered and that arising from the possibility that the model does not accurately represent the phenomena under study (Bostrom & Ćirković, 2008).

These uncertainties are often difficult to analyze using rigorous statistical methods. Even analysts themselves sometimes tend to exclude variables that they cannot objectively quantify. However, this omission can lead to significant errors in judgment. Taylor argues that the largest deviations are typically found at the extremes of the probability of exceedance curves, which leads to an underestimation of the risk associated with certain extreme events.

The author also examines two risk perception studies conducted among business leaders. The first was conducted by the insurance company Swiss Re in 2005 and surveyed executives at multinational companies about the financial threats that concerned them most. The results show that, from the perspective of individual companies, global catastrophic risks took a back seat to more immediate business threats.

The second study was conducted by the World Economic Forum. Its Global Risks 2007 report ranked various hazards according to their probability and severity, based on the opinions of business leaders, political figures, economists, and academics. This type of exercise allows us to observe how different institutions prioritize threats, although it can also reproduce biases stemming from the interests and time horizons of those participating in the assessment.

Chapter 9, written by legal scholar and economist Richard Posner, addresses the challenges that public policy faces in managing global catastrophic risks. Posner argues that government action is limited by the short-term time horizons of policymakers, the restricted duration of their terms in office, and the large number of immediate problems vying for their attention. Furthermore, mitigating catastrophic risks often requires substantial investments, making it difficult to allocate sufficient budgets.

Posner analyzes public policies aimed at specific risks, including tsunamis, asteroid impacts, bioterrorism, particle accelerator experiments, and global warming. Although it is not always possible to make completely rigorous estimates, he considers it necessary to attempt to quantify the probabilities, potential damages, and costs of preventive measures. The purpose of this exercise is to establish priorities and design more effective mitigation strategies.

When it is not possible to assign a precise probability to a threat, Posner proposes using the concept of implicit probability. This can be inferred by comparing the resources allocated to mitigation with the magnitude of the losses that the catastrophe would produce. For example, if one million dollars is allocated annually to reduce a risk capable of causing one billion dollars in damages, it can be interpreted that the current policy implicitly assumes an annual probability close to 1 in 1,000 (Bostrom & Ćirković, 2008).

However, this calculation does not necessarily mean that policymakers have carried out a conscious or adequate assessment. It may also reveal that it has been decided, explicitly or implicitly, that one million dollars constitutes a sufficient investment, even if the actual probability of the disaster could be greater than 1 in 1,000. Posner’s central criticism, therefore, is that governments allocate insufficient resources to the prevention and mitigation of global catastrophic risks.

Part II. Natural Hazards

Regarding natural hazards, volcanic eruptions recorded in recent history have produced measurable effects on the global climate. In some cases, they have caused a temporary cooling of the Earth by several tenths of a degree, lasting approximately one year. However, as Michael Rampino points out in Chapter 10, these episodes are small compared to the largest known eruptions.

One of the most significant cases is the eruption of the Toba supervolcano, which occurred in Indonesia approximately 75,000 years ago. This event ejected enormous quantities of fine ash and aerosols into the atmosphere, generating climatic effects comparable to those of a nuclear winter.

According to some estimates, the human population then experienced a drastic reduction for at least one generation. Some hypotheses suggest that it may have decreased to around 4,000 individuals, among whom there would have been approximately 500 women of reproductive age. According to the Toba catastrophe theory, this population decline was a direct consequence of the supereruption and brought the human species to the brink of extinction. Rampino considers this event one of the most serious catastrophes humanity has ever faced.

The main global danger of supervolcanism lies in its climatic consequences. A large-scale eruption could trigger a volcanic winter, characterized by a prolonged drop in temperatures, a reduction in solar radiation, and a significant decline in agricultural productivity. These effects could lead to widespread famines, economic crises, and severe social unrest.

Rampino’s analysis is also relevant for understanding other risks, such as nuclear war or the impact of an asteroid or meteorite. Although these events have different causes, they could all release large quantities of soot, dust, and aerosols into the atmosphere, reducing the amount of solar radiation reaching the Earth and causing global cooling.

While it is not possible to prevent a super-eruption, measures can be taken to mitigate its consequences. One such measure would be to expand strategic food reserves. Currently, global grain stocks can feed the population for a relatively short period, approximately two months. However, a catastrophe of this nature could reduce agricultural production for several years. Therefore, establishing larger and more diversified reserves would help prevent famines, both in supervolcanic scenarios and in other situations capable of causing a temporary decrease in global food production.

In Chapter 11, William Napier examines the catastrophic risks associated with comets, asteroids, and meteorites. The author explains the available scientific knowledge about these objects: their origin, how frequently they approach Earth, and the potential consequences of an impact.

Napier argues that the level of risk associated with these phenomena could be compared to that of air accidents and other widely recognized threats. However, there is a significant difference: the resources allocated to preventing and mitigating cosmic impacts are negligible compared to the funding assigned to aviation safety. To address this disparity, the Spaceguard project has been launched, aimed at detecting, identifying, and monitoring potentially hazardous near-Earth objects.

Because asteroids constitute the majority of the near-Earth object threats, one of the main strategies involves detecting them early enough to alter their trajectory. Several studies have evaluated the possibility of deflecting an asteroid using technologies capable of changing its course before impact. Manufacturing and deploying these defense systems would be more expensive than simply searching for and monitoring potential impactors. Nevertheless, their cost could be considered reasonable given that their purpose would be to protect civilization and, potentially, all life on Earth.

Asteroids and comets, however, are not the only hazards from space. There are other cosmic threats, including climate changes resulting from fluctuations in solar activity, large radiation flows and cosmic rays produced by supernova explosions or gamma ray bursts (Bostrom & Ćirković, 2008).

These risks are analyzed by Arnon Dar in Chapter 12. Although his study is relevant to understanding the external threats facing Earth, the possibilities for mitigation are still very limited. In some cases, the hazards seem extremely remote; in others, humanity lacks the knowledge and technological resources necessary to prevent them or reduce their effects. For this reason, research remains the primary tool available to assess their probability, understand their mechanisms, and anticipate their potential consequences.

Part III. Risks Arising from Unintended Consequences

Despite the diversity of threats examined up to this point, global warming caused by greenhouse gas emissions has garnered a considerable share of public attention. Anthropogenic climate change has, in fact, become the paradigmatic example of a global threat. According to Bostrom and Ćirković, this issue occupies a disproportionate share of the attention given to all global catastrophic risks (Bostrom & Ćirković, 2008).

The accumulation of carbon dioxide and other greenhouse gases in the atmosphere causes an increase in the planet’s average temperature and contributes to sea level rise. In this regard, reports from the United Nations Intergovernmental Panel on Climate Change (IPCC) estimate the increase in global temperature that could occur by the end of the 21st century in the absence of effective mitigation measures.

These projections are subject to varying degrees of uncertainty. Among the most important factors are the future evolution of emissions, the true value of climate sensitivity, and the possible interactions between different components of the Earth system. For this reason, the IPCC presents several climate scenarios constructed from different models and assumptions about humanity’s economic, technological, and energy development.

In Chapter 13, David Frame and Myles Allen explain the basic principles of climate modeling and pay particular attention to low-probability scenarios with extreme consequences. Although these outcomes may seem unlikely, they are precisely the ones that generate the greatest concern due to the magnitude of their potential effects and the uncertainty surrounding their estimation.

Frame and Allen also examine the political difficulties associated with mitigation. One of the main problems is determining which targets to adopt when there is no certainty about the exact level of emissions capable of causing dangerous anthropogenic interference with the climate system. This uncertainty complicates both the definition of thresholds and the distribution of responsibilities and costs among different countries.

In Chapter 14, the American virologist Edwin Kilbourne analyzes some of the most significant pandemics in history, the characteristics of their pathogens, and the factors that could contribute to a large-scale catastrophe. He also examines the potential consequences of future outbreaks. As the authors remind us, infectious diseases have caused enormous amounts of suffering and death throughout human history and continue to do so today (Bostrom & Ćirković, 2008).

Pandemics constitute one of the major contemporary global catastrophic risks, although they are often not perceived as such. An example of this underestimation is the tendency to consider the 1918–1919 influenza pandemic as a less serious event than the two world wars. However, available estimates show that its demographic impact was comparable to, and even greater than, that of some of those conflicts.

In addition to combating diseases that already affect the world’s population, it is essential to prevent the emergence and spread of new outbreaks with pandemic potential, such as severe acute respiratory syndrome, avian influenza, or drug-resistant variants of tuberculosis. Early and coordinated intervention can, under certain circumstances, halt an emerging pandemic before it spreads and save millions of lives (Bostrom & Ćirković, 2008).

The authors place pandemics within the category of risks stemming from unintended consequences, even though most infectious agents have a natural origin. The reason is that the evolution, spread, and destructive capacity of pathogens increasingly depend on conditions created by human civilization. The global spread of diseases, for example, became possible with the progressive interconnection of transportation routes between inhabited continents.

Currently, the intensification of travel, trade, and the international circulation of goods allows a contagious disease to spread across the globe in a matter of days or weeks. Kilbourne also identifies another factor related to globalization: the homogenization of populations, practices, and cultures. The more a human population resembles a single homogeneous niche, the greater the likelihood that a single pathogen will spread rapidly throughout the system (Bostrom & Ćirković, 2008).

Based on this idea, Kilbourne refers to the so-called “bad apple syndrome,” linked to the mass production of food and the spread of uniform behavioral habits. A single contaminated product can reach millions of consumers if it is part of a large-scale industrial production and distribution chain. Under these conditions, a localized error ceases to be a minor incident and can become an international problem.

Cultural, genetic, and productive diversity, on the other hand, acts as a form of protection. The greater the variety of practices and systems, the less likely a single pattern will spread universally before its risks are identified. This reasoning can be applied not only to pathogens but also to hazardous chemicals, flawed technologies, or even oppressive ideologies.

Unlike pandemics, artificial intelligence does not yet represent an ongoing catastrophic process or a clearly observable global risk. However, from a long-term perspective, the development of artificial intelligence capable of far surpassing human cognitive abilities could become one of the main challenges for the future of humanity. In Chapter 15, Eliezer Yudkowsky analyzes AI both as a potential source of extraordinary benefits and as a possible catastrophic threat.

Yudkowsky begins by clarifying several terms and correcting some misconceptions about artificial intelligence. From this, he argues that radical superintelligence could emerge through a process of rapid self-optimization associated with the so-called singularity hypothesis. According to this view, a distinction must be made between when the transition to superior intelligence begins and the speed at which this transformation could unfold (Bostrom & Ćirković, 2008).

It is possible that computers will take a long time to achieve capabilities comparable to those of human thought. However, once this threshold is reached, they could rapidly surpass human intelligence. Yudkowsky conceives of superintelligence as an extremely powerful optimization system, capable of transforming the world to achieve its assigned objectives. To avoid catastrophic results, it would be essential to ensure that such a system pursues goals compatible with human well-being.

The author identifies two main ways in which the construction of a friendly artificial intelligence could fail. The first is philosophical in nature: the system could be assigned a seemingly reasonable goal, but one that leads to morally unacceptable consequences or consequences radically different from those desired. The second is technical: even if the goal were appropriate, its implementation could be flawed or produce unexpected effects due to errors in how it is specified.

For this reason, Yudkowsky considers it necessary to investigate the foundations of an artificial intelligence aligned with human values ​​before attempting to build a superintelligence. The issue is not merely about increasing the capabilities of systems, but about ensuring that their goals, methods, and decision-making processes do not conflict with the survival and well-being of humanity.

Chapter 16 examines the possibility that certain experiments conducted in particle accelerators could generate an existential risk. In response to concerns surrounding these experiments, the director of the Relativistic Heavy Ion Collider at Brookhaven National Laboratory commissioned an official safety report in 2000. Similar concerns later arose regarding the Large Hadron Collider.

Following this debate, the physicist Theoretical scientist Frank Wilczek considers three possible catastrophic scenarios. The first would be the formation of small black holes capable of accumulating surrounding matter until they consume the Earth. The second would consist of the production of stable, negatively charged strangelets, capable of catalyzing the transformation of ordinary matter into strange matter. The third would be the triggering of a vacuum transition that would propagate at the speed of light and destroy not only the planet, but the entire causally accessible region of the universe (Bostrom & Ćirković, 2008).

Wilczek considers these scenarios extraordinarily improbable for both theoretical and empirical reasons. Cosmic rays reach energies higher than those produced in human-made accelerators and have collided with Earth’s atmosphere for billions of years without catastrophic consequences. If artificial collisions do not differ significantly from those that occur naturally, this evidence constitutes a strong argument in favor of the safety of accelerators.

However, the issue is not simply a matter of determining whether a catastrophic outcome is improbable. The possibility that the models, assumptions, or calculations used to assess the risk are incorrect must also be considered. Therefore, it is necessary to avoid overconfidence bias and recognize that even specialists can make mistakes.

Physicists possess the technical expertise to assess the risks of high-energy experiments, but they are part of professional communities with a vested interest in the continuation of such research. This can generate, at least from the public’s perspective, the suspicion that there are incentives to downplay the dangers. Conversely, some experts can also be influenced by media pressure and exaggerate risks that lack a solid scientific basis. The central problem, therefore, is determining who should assess these threats, using what procedures, and under what conditions of independence.

In Chapter 17, economist Robin Hanson studies societal collapse as a mechanism capable of amplifying other catastrophes. To illustrate his argument, he uses the example of a fall down a flight of stairs. The main danger lies not in slipping on a single step, but in that this first stumble triggering a series of falls and ultimately causing serious injuries. Similarly, catastrophes should not be analyzed solely in terms of their direct damage, but also in their capacity to destabilize economic, institutional, and social systems.

This reasoning does not apply equally to all risks. In scenarios such as the creation of a hostile superintelligence or a physical catastrophe generated by a particle accelerator, there may be little possibility of intermediate outcomes. In contrast, phenomena such as storms, earthquakes, floods, forest fires, terrorist attacks, plagues, or wars exhibit a wide range of intensities. In these cases, a considerable portion of the potential damage stems from the social disruption and collapse processes that can be triggered.

Hanson argues that many of these risks follow a power-law distribution. This means that most of the expected damage may be concentrated in a large number of small events or, conversely, in a small number of extreme catastrophes. Car accidents, for example, exhibit a high exponent: most deaths result from numerous small-scale accidents. Wars and pandemics, on the other hand, show lower exponents, as a considerable proportion of the damage is concentrated in a few conflicts or outbreaks of enormous magnitude.

Hanson also applies the theory of economic growth to the analysis of its reverse: the accelerated decline in productivity caused by the destruction of social, institutional, and material capital. A deep crisis can deteriorate the cooperation and coordination mechanisms on which the economy depends.

For example, a judge who would normally reject a bribe might accept it if his life or family were threatened. Such corruption would encourage new crimes, reduce trust in institutions and discourage investment. Likewise, a widespread loss of confidence in banks could lead to massive withdrawals of funds and make the financial system difficult to operate. A relatively limited initial shock could thus trigger a spiral of institutional deterioration.

The productivity of the global economy depends both on its scale and on the precise coordination between multiple forms of capital (Bostrom & Ćirković, 2008). This interdependence creates the risk that a localized shock will affect a particularly fragile component and spread to cause a far-reaching economic collapse.

Finally, Hanson proposes a possible strategy to transform certain existential risks into non-existential risks. It proposes considering the construction of self-sufficient shelters, possibly located in underground facilities and equipped with food, tools and other essential resources. These spaces could house a small group of survivors capable of preserving knowledge and contributing to the reconstruction of a post-catastrophe society, in a kind of technological “Noah’s Ark.”

Disaster prevention remains a preferable alternative to any subsequent survival strategy. However, Hanson considers it pertinent to evaluate whether any variant of these refuges could be viable as a complementary measure to reduce the probability of human extinction in extreme scenarios.

Part IV. Risks Arising from Hostile Acts

Regarding the risks arising from hostile acts, the threat of a nuclear Armageddon that dominated the public imagination during the Cold War seems to have lost some of its prominence. This is largely due to the emergence of other global concerns, such as the environmental crisis and terrorism. However, Chapter 18 warns that nuclear war continues to pose a serious threat. Relations between the United States and Russia could deteriorate again to the point where a political or military crisis triggers a nuclear confrontation. Furthermore, a new arms race could lead to an increase in the arsenals of both countries.

Another possibility is that a nuclear war could occur between powers other than the former Cold War adversaries. This danger would increase as new countries join the so-called nuclear club, especially those involved in persistent regional conflicts, such as India, Pakistan, North Korea, and Israel. The more states that possess nuclear weapons, the more difficult it will be to contain proliferation. The dissemination of knowledge and technological capabilities would progressively reduce technical barriers, while some countries that previously renounced this type of weaponry might feel compelled to reconsider their decision and return to the nuclear path.

A third possibility is the accidental start of a nuclear war. Joseph Cirincione recalls an incident in 1995 when Russian military forces mistook a Norwegian weather rocket for a ballistic missile launched from a U.S. submarine. Then-President Boris Yeltsin activated the portable nuclear command system for the first time and had only a few minutes to decide whether to order a counterattack. Ultimately, radar confirmed it was a false alarm, and no launch took place.

Other events have also brought the world close to a nuclear confrontation. The best-known example is the Cuban Missile Crisis, during which, according to estimates attributed to President John F. Kennedy, there was a considerable probability that the confrontation between the United States and the Soviet Union would escalate into a nuclear war. To reduce these dangers, Cirincione proposes resolving regional conflicts, strengthening the Treaty on the Non-Proliferation of Nuclear Weapons, and making progressive progress toward the elimination of nuclear arsenals.

In Chapter 19, William Potter and Gary Ackerman analyze the various forms that nuclear terrorism could take. These include the dispersal of radioactive material using conventional explosives, the sabotage of nuclear facilities, the acquisition of fissile material to manufacture and detonate a rudimentary device, the acquisition and use of an already constructed nuclear weapon, and the use of deception tactics designed to provoke a state into launching a nuclear warhead (Bostrom & Ćirković, 2008).

Potter and Ackerman focus their attention on high-consequence nuclear terrorism, a category that primarily encompasses the last three possibilities. The authors examine both the supply and demand for nuclear materials and capabilities, the potential consequences of an attack, the future evolution of the threat, and the policies necessary to prevent it.

Up to the point analyzed in this work, no non-state actor had managed to acquire a fission weapon. Nor was there conclusive evidence that organizations such as Al-Qaeda or Aum Shinrikyo had succeeded in transforming their motivations, economic resources, organizational capabilities, support networks, and potential links with allied states into a genuine nuclear capability (Bostrom & Ćirković, 2008). Based on the limited publicly available information, it can be inferred that the main obstacle was access to the fissile material necessary to construct the weapon.

Despite this lack of precedent, some specialists continued to consider the risk high. Graham Allison, for example, estimated a near 50% probability of a nuclear attack occurring within the next decade if radical nonproliferation measures were not adopted. Other experts offered considerably lower estimates, revealing the difficulty of assigning precise probabilities to such threats.

Overall, the analysis underscores the importance of preventing nuclear terrorism and, in particular, restricting non-state actors’ access to fissile materials (Crucial Fact 24). Potter and Ackerman argue, however, that the scarcest resource is not necessarily technical or financial, but rather the sustained, high-level political leadership needed to transform public pronouncements into effective policies.

In Chapter 20, astrobiologist Christopher Chyba and molecular biologist Ali Nouri examine the issues related to biotechnology and biosecurity. Both nuclear and biotechnology technologies can be used to manufacture weapons of mass destruction, although there are important differences between them. First, biological weapons can be developed in small, easily concealed facilities without requiring particularly unusual raw materials. Second, an infectious agent can spread far beyond its point of release and potentially reach the entire planet.

Biosecurity threats can be divided into several categories: naturally occurring diseases, illicit state biological weapons programs, actions by non-state actors, biohacking activities, laboratory accidents, and unintentional releases of pathogens (Bostrom & Ćirković, 2008). Historically, naturally occurring diseases have caused several orders of magnitude more deaths than deliberate or accidental threats associated with biotechnology. However, the dangers posed by biotechnology could increase as technical capabilities expand and become more widespread.

The authors present several recent developments that illustrate this concern. First, a group of Australian researchers attempted to control the rabbit population by introducing the interleukin-4 gene into a mousepox virus, with the aim of sterilizing the animals. The result was unexpected: the virus suppressed the host’s immune response and caused the death of all infected animals, including those that had been vaccinated. Subsequently, another team developed a variant that proved completely lethal in vaccinated mice, even when they received antiviral treatment.

Second, the poliovirus was synthesized using commercially available chemical materials. The first time this procedure was performed, it required a complex and lengthy research project. Since then, the time required to synthesize viral genomes of similar size has been considerably reduced. The virus responsible for the 1918 influenza pandemic was also reconstructed and is now preserved in laboratories in the United States and Canada.

Third, technologies designed to modify the properties of viruses and other microorganisms are progressing rapidly. RNA interference allows for the deactivation of specific genes in humans and other organisms, while synthetic biology is establishing itself as a field focused on the design of biological devices and new classes of microorganisms.

The public availability of the complete genomes of hundreds of bacteria, fungi, and viruses increases research possibilities but also generates significant risks. As technical barriers decrease and biotechnological knowledge, equipment, and materials become more widespread, the creation of pathogens with particularly dangerous characteristics becomes more feasible.

The dual-use nature of equipment and knowledge, along with the possibility of operating in small and difficult-to-detect facilities, poses considerable challenges for regulation. An effective regime must strike a balance between preventing abuses and enabling the research necessary to develop treatments, vaccines, and diagnostic methods. Chyba and Nouri examine various strategies for strengthening biosafety, including the automated analysis of genetic sequences sent to centralized DNA synthesis facilities (Bostrom & Ćirković, 2008). All indications are that this area will acquire increasing importance and will require a multifaceted approach capable of responding to the possible deliberate creation of new pathogens.

In Chapter 21, Chris Phoenix and Mike Treder study nanotechnology as a potential source of global catastrophic risks. The authors distinguish between existing or soon-to-be-developed nanoscale technologies and molecular manufacturing, which remains a future technological possibility. The former do not appear to pose new catastrophic risks in themselves, although they could amplify or mitigate other threats. Phoenix and Treder therefore focus their analysis on the capabilities and dangers associated with molecular manufacturing. Although this technology did not constitute an immediate risk, in the long term it could become an extremely serious threat.

Molecular nanotechnology would considerably expand human control over the structure of matter. Molecular machinery systems would allow for the rapid and inexpensive manufacture of microscopic and macroscopic objects with atomic precision. These production platforms could be composed of millions of microscopic assembly mechanisms, capable of building objects through the controlled addition of molecules to a structure.

The possibilities of this technology would far surpass the capabilities of molecular assemblers found in nature. Its potential applications include nanofactories, medical nanorobots, high-speed computers, lightweight and strong diamond materials, systems for removing pollutants, home manufacturing plants capable of producing objects from downloadable blueprints, low-cost solar collectors, advanced space technologies, and mass-produced sensors.

However, these same capabilities could be used to economically produce improved conventional weapons and new classes of armaments. Molecular manufacturing would thus have both beneficial and destructive applications.

Phoenix and Treder identify several catastrophic risks associated with this technology: armed conflicts, economic and social collapse, authoritarian forms of global governance, uncontrolled acceleration of artificial intelligence development, environmental damage, and the release of nanomachines capable of consuming or destroying the biosphere—a scenario known as ecophagy.

In the absence of effective preventative and defensive systems, molecular manufacturing could grant destructive capabilities to a wide range of actors, including individuals, groups, corporations, and states. Even if the likelihood of a powerful actor deliberately deciding to exterminate humanity were reduced, the risk of the accidental development or release of devastating weapons should also be considered.

Likewise, a conflict between actors with comparable destructive capabilities could escalate until one of the parties felt compelled to employ weapons with apocalyptic consequences. Therefore, until reliable defenses against this type of threat exist, it is a priority to limit the number of actors capable of acquiring molecularly manufactured weapons technologies and to develop control mechanisms before these capabilities become widespread.

In Chapter 23, Bryan Caplan analyzes another catastrophic global risk: totalitarianism. The American economist reminds us that the regimes of Nazi Germany, Stalinist Soviet Union, and Maoist China were responsible for tens of millions of deaths during the 20th century. Unlike threats such as an asteroid impact, totalitarianism is difficult to analyze completely objectively due to ideological differences regarding its definition, causes, and the measures needed to prevent it. Nevertheless, extreme forms of political oppression must be considered a serious risk, as they have been one of the most persistent threats in human history and remain a real possibility.

Caplan points out that totalitarianism not only poses a danger in itself, but can also exacerbate other risks. In these regimes, officials and citizens often fear communicating bad news, while leaders restrict criticism and dissenting opinions. As a result, the political system may ignore emerging threats and make serious decision errors.

Nevertheless, totalitarian regimes possess certain operational advantages over democratic societies. Their capacity to employ force, suppress opposition, and massively mobilize resources can make them effective in achieving specific objectives. This effectiveness contributes, at least temporarily, to their stability.

Caplan identifies two factors that have historically limited the duration of these systems. The first is the problem of succession. A charismatic or authoritarian leader may retain power throughout their life, but the stability of the regime can weaken when a successor capable of maintaining the status quo and controlling the various political factions must be appointed.

The second factor is the existence of non-totalitarian countries. Open societies offer inhabitants of authoritarian regimes alternative examples of freedom and prosperity, which can increase their dissatisfaction with the system. To avoid this influence, leaders may restrict contact with the outside world, as happened in communist Albania and continues to happen in North Korea.

However, no isolation is completely airtight, and information can leak out. Moreover, an excessively isolated country risks falling behind economically and militarily, becoming vulnerable to invasion or regime change imposed from abroad.

Future technological developments could weaken these restrictions. Advances in surveillance would allow for more precise control of the population, while certain neurological technologies could be used to identify or indoctrinate hidden opponents. Drugs designed to increase docility without reducing productivity could also be developed.

Life-extension medicine could also extend a single leader’s rule for decades or centuries, reducing the problem of succession. Regarding the limitations imposed by the existence of free countries, Caplan expresses concern about the possibility of a world government. Even if it begins as a democratic institution, it could later transform into a totalitarian regime. A totalitarianism of global reach would be especially difficult to overthrow, as it would lack external adversaries capable of containing it or offering alternative political models.

To discuss these issues productively, Caplan argues that it is necessary to compare the benefits and risks of each measure, avoiding analyzing a single risk in isolation. For example, indefinitely extending lifespan might slightly increase the likelihood of a stable totalitarian regime emerging. However, it would be disproportionate to force the entire population to die of old age to avoid a remote possibility of falling victim to a secret police force a thousand years from now (Bostrom & Ćirković, 2008).

The conclusion is not to reject any technological advance or form of global governance, but rather to design them with their potential authoritarian effects in mind. It is possible to strengthen new international institutions while simultaneously establishing controls designed to reduce the risk of their evolving into a global totalitarian regime.

Conclusions and Future Directions

In conclusion, Bostrom and Ćirković argue that the most likely global catastrophic risks originate from human activities, particularly industrial civilization and the development of advanced technologies. However, this does not mean that industry and technology should be considered solely responsible for such threats, as they have also generated significant benefits for humanity. While technological modernity has given rise to new catastrophic risks, it has also contributed to reducing numerous smaller-scale hazards in different parts of the world. In fact, local and personal disasters, such as famine, thirst, predation, disease, and small-scale violence, have historically caused far more deaths than major global cataclysms (Bostrom & Ćirković, 2008).

Considering available statistical trends, the contemporary world can be described, in general terms, as a safer place than in previous eras. Global catastrophic risks are the most serious in terms of their potential scope, but not necessarily in terms of expected damage. At the same time, technology and complex forms of social organization provide important resources for mitigating many existing threats. Nevertheless, it is crucial to recognize that the greatest risks facing humanity are not entirely external, but are closely linked to the direct and indirect, foreseen and unforeseen, consequences of its own behavior.

One of the major contemporary catastrophic risks is the emergence of a large-scale infectious pandemic. However, mortality statistics pose difficulties in determining which diseases should be included in this category. If the analysis is conducted at a more precise level of disaggregation and specific infectious agents are considered, some ongoing pandemics could already be classified as global catastrophes. HIV/AIDS, for example, was causing approximately three million deaths annually at the time of publication (Bostrom & Ćirković, 2008). By similar reasoning, cardiovascular diseases and various types of cancer could also be understood as persistent global catastrophes (Crucial Fact 26).

It is also necessary to consider aging, one of the main underlying causes of death and disability. This process is linked to a considerable proportion of the approximately 57 million deaths recorded annually, in addition to producing a significant loss of health and human capital. If aging were an uncertain possibility, rather than a virtually universal condition, it would likely rank among the top factors in any classification of catastrophic risks.

The fact that aging is a common and seemingly inevitable cause should not lead to underestimating its importance. If realistic means existed to delay its effects, through the dissemination of healthy lifestyle habits or substantial investment in biogerontological research, the expected number of deaths could be substantially reduced. From this perspective, partial progress on a problem of enormous magnitude could yield greater benefits than the complete elimination of certain far less likely catastrophic risks.

Other risks are more clearly catastrophic, including nuclear war, nuclear terrorism, and the destructive use of biotechnology. Over longer time horizons, the dangers associated with molecular manufacturing, artificial intelligence, and totalitarianism could become increasingly significant. Each of these could even constitute an existential risk (Bostrom & Ćirković, 2008).

However, the fact that one risk is greater than another does not necessarily imply that more resources should be allocated to its management. Some threats may be virtually impossible to mitigate, while in other cases, preventive measures could be excessively costly or even generate new dangers. Conversely, a relatively small risk may deserve priority attention when its solution is economical, accessible, and easy to implement.

Anthropogenic ozone depletion is an example of a serious problem whose mitigation has progressed considerably thanks to a coordinated international response. Nevertheless, there are other, more urgent and less controlled risks that require immediate attention. Resource allocation should therefore be based not only on the magnitude of each threat, but also on its probability, its mitigation capacity, and the relationship between intervention costs and expected benefits.

The authors consider it essential to continue investigating individual long-term risks, especially those that are potentially more serious and still poorly understood, such as those associated with biotechnology, molecular manufacturing, artificial intelligence, and systemic vulnerabilities. It is also necessary to identify, compare, and evaluate potential mitigation strategies.

In some cases, technically feasible and economically accessible countermeasures already exist, but their implementation depends on the presence of political and institutional leadership capable of transforming them into effective programs. Alongside these practical efforts, research is needed to clarify the conceptual and methodological problems that arise when studying threats with very low probability and extreme consequences.

To conclude the introduction, Bostrom and Ćirković propose several courses of action aimed at strengthening the study and management of global catastrophic risks. First, they suggest directing some research toward producing information useful for decision-making. This includes identifying early warning signs, creating indicators to measure progress in risk reduction, and developing qualitative and quantitative assessment models.

Secondly, they recommend designing and implementing more advanced methodologies to integrate data, expert opinions, and probabilistic forecasts. These tools include prediction markets and other mechanisms designed to gather disparate information and improve the quality of estimates.

Thirdly, they consider it necessary to pay greater attention to the development and evaluation of specific mitigation strategies. These should be analyzed both for their direct usefulness and for the political, economic, and institutional instruments needed to implement them. The goal is not only to understand the risks but also to determine which interventions can effectively reduce them.

The authors also emphasize the need to give special attention to existential risks and the particular methodological problems involved in studying them. Due to the magnitude and irreversibility of their potential consequences, these threats cannot be assessed solely using the same criteria employed for ordinary risks.

Another priority is to build a stronger, interdisciplinary, and international community, comprised not only of specialists from diverse academic fields, but also of practitioners, policymakers, and institutional actors responsible for implementing risk reduction measures. Such a community would help bridge the gap between theoretical research and practical intervention.

Finally, Bostrom and Ćirković consider it essential to promote a more critical, reflective, and analytical public discourse on global catastrophic risks. Their study and management must be situated within a broader framework of challenges and opportunities aimed not only at preventing destruction but also at safeguarding and improving the human condition.

Data on Timescales and Existential Risks

1. Cosmic and Planetary Timescales

2. Mass Extinctions

• Frequency: 15 mass extinctions over the last 500 million years; 5 of them eliminated more than half of all existing species.
• The “Great Dying” (251.4 million years ago): the largest known extinction event. It wiped out more than 90% of species and entire phylogenetic families. Biodiversity required more than 5 million years to recover.
• The impact 65 million years ago (the extinction of the dinosaurs): possibly a necessary condition for the rise of Homo sapiens by freeing the ecological niche previously occupied by large reptiles.
• Overall estimate: at least 99.9% of all species that have ever existed are now extinct.

3. Human Evolution

4. Global Risk Assessment (World Economic Forum, 2007 Report, 10-Year Horizon)

By priority/severity:

1. Computational risk (highest priority)
2. International trade
3. Corporate governance
4. Operational risks
5. Infrastructure risks
6. Liability risks
7. Natural disasters
8. (Not specified)
9. Terrorism

By economic losses (>USD 1 trillion over 10 years):

• Asset price collapse: 10–20% probability
• Decline in globalization: 1–5% probability

By expected fatalities:

1. Pandemics
2. Diseases in the developing world
3. Civil and interstate wars

Note: several risks included in the survey were poorly defined, making the results difficult to interpret. Clearly defining risks is a necessary prerequisite for meaningful quantification.

5. Catastrophic Climate Impacts (Nuclear or Impact Winter)

• Estimated global cooling: −15 °C
• Ocean surface cooling: 2–6 °C, potentially lasting several years
• Persistent atmospheric soot (1–3 years) could trigger climatic cooling lasting decades, amplified by positive feedback mechanisms (greater snow and ice cover → higher solar reflectivity)

6. Asteroid and Comet Impact Risk

• Volcanic super-eruptions: occur on average every 50,000 years (possibly underestimated because of incomplete geological records).
• NASA’s Spaceguard Project (~USD 4 million annually plus volunteer support): aimed to detect 90% of near-Earth asteroids larger than 1 km by the end of 2008.

7. Climate Change (IPCC Projections)

• Sea-level rise (under the scenarios considered): 18–38 cm and 26–59 cm

8. Infectious Diseases and Global Mortality

• Account for approximately 25% of all deaths worldwide (~15 million annually)
• Approximately 75% of these deaths occur in Southeast Asia and Sub-Saharan Africa.

Five leading infectious causes of death:

1. Respiratory infections — 3.9 million deaths/year
2. HIV/AIDS — 2.9 million deaths/year
3. Diarrhoeal diseases — 1.8 million deaths/year
4. Tuberculosis — 1.7 million deaths/year
5. Malaria — 1.3 million deaths/year

Historical comparison: World War I vs. the Spanish Flu

• World War I: ~10 million military deaths + 9 million civilian deaths
• Spanish Flu: 20–50 million deaths (despite a case fatality rate of only 2–3%; the exceptionally high death toll resulted from the enormous proportion of the global population infected)

Non-communicable diseases:

• Cardiovascular disease: ~30% of global mortality (~18 million deaths/year)
• Cancer: ~8 million deaths/year

9. Nuclear Weapons

• Global nuclear arsenal: from 65,000 warheads (Cold War peak, 1986) to ~26,000 (2007), representing a 60% reduction
• Approximately 96% of these weapons were held by the United States and Russia.
• Global plutonium stockpile: ~2,000 tonnes (around ten times the amount contained in active warheads)

Scenario of full-scale U.S.–Russia nuclear war:

• United States: 35–77% of the population killed (105–230 million)
• Russia: 20–40% of the population killed (28–56 million)
• Indirect effects (economic collapse and a possible nuclear winter) could substantially increase these figures.

Prevention initiative:

• G8 Global Partnership (2002): commitment to invest USD 20 billion over ten years to prevent terrorists from acquiring weapons of mass destruction.

Citation

Bostrom, Nick, and Milan Ćirković (2008), Global Catastrophic Risks, Oxford: Oxford University Press.