Long-term risks of nuclear weapons additional to nuclear winter

The EA community has reinvigorated research on the probability that a nuclear war could cause a nuclear winter that might cause human extinction.  Beyond this important question about the climate effects of a single nuclear war, nuclear weapons threaten humanity with additional important, neglected, and tractable existential risk factors.  These additional, interrelated nuclear weapons existential risk factors include: 1) repeated nuclear wars, 2) a single point of failure for planetary-scale destruction, 3) distributed capacity for city-scale destruction, and 4) changed geology permanently favoring human access to nuclear weapons.  Each of these risk factors and potential responses are discussed briefly below.

1. Repeated nuclear wars

The risk of repeated nuclear wars is existentially important.  Assuming a low risk of nuclear war unduly privileges the present. If low annual risks of nuclear war stack indefinitely, eventually they will rise above unity and become a frequency of future nuclear wars.  As mathematician and physicist Freeman Dyson explained in his 1984 book Weapons and Hope: “although it may be technically impossible for a single nuclear exchange to exterminate the human species, a succession of nuclear wars might well do so…If a species becomes collectively insane, then it becomes in the Darwinian sense unfit, and in the long run it is unlikely to survive.  A sequence of ten nuclear wars, each one more desperate and more insane than the one before, could plausibly result in our final disappearance from the planet.”  Nuclear arsenals are the only components of our civilization designed to “survive” nuclear warfighting – governments; physical infrastructure for electricity, water, transportation, and health care; and social cohesion are not. Humanity’s fragile means for preventing nuclear war through deterrence and diplomacy are extremely vulnerable to any use of nuclear weapons.  Today’s human capacity to prevent future nuclear wars would likely be destroyed by the next nuclear war.

The risk of repeated of nuclear wars is neglected because ensuring a nuclear war does not end humanity’s capacity to prevent nuclear war requires acknowledging the contingency of today’s governments.  Today’s nuclear armed states seek to deter adversaries by posturing for nuclear warfighting.  Governments plan to destroy adversary militaries, governments, and societies through nuclear attack, but do not have the means to defend themselves against large-scale nuclear attack.  The mechanism driving government neglect of this risk is illustrated by the case of RAND Corporation analyst Paul Kesckemeti whose 1958 book Strategic Surrender sparked furious debate in the U.S. Congress leading to a ban on the use of government funds to consider U.S. surrender.  Governments and militaries plan to win wars or be destroyed, not for what comes after their destruction.

The risk of repeated nuclear wars is tractable.  The fact that today’s governments cannot address this risk does not prevent others from doing so.  Deterrence theory, strategic stability, arms control, threat reduction, and nuclear security are examples of ideas that originated outside government that each revolutionized nuclear weapons policy.  Systems of nuclear restraint that could survive a nuclear war will require political, institutional, and technological innovations built to survive the destruction of one or more governments and grievous damage to human civilization.  It is easy to imagine a version of the Svalbard Global Seed Vault that, instead of seeds, stores planetary capacity to restore control over nuclear weapons after a nuclear war devastates today’s governments; but it isn’t anyone’s job to imagine what comes after the United States, the Russian Federation, and the People’s Republic of China.  Important conceptual work on this problem could begin for less than $100,000.

2. Single point of failure for planetary-scale destruction

The risk of having a single point of failure for planetary-scale destruction is existentially important, even if this destruction does not directly exterminate humanity.  Former nuclear war planner Daniel Ellsberg frames the risk of nuclear war as the product of continuous interaction among various parts of a single, global, “omnicidal” system in his 2017 book The Doomsday Machine.  The “Doomsday Machine” comprises nuclear arsenals, other military forces, and the human fears and uncertainties that lead to these material components of the system.  While activating Ellsberg’s “Doomsday Machine” might not prove “omnicidal,” it is designed to kill hundreds of millions immediately and, through climate effects, might kill a billion or more. 

The risk of a single point of failure for planetary-scale destruction is neglected.  Governments and most analysts understand the world’s nuclear arsenals as the sovereign property of nine independent governments. Each of these nine governments holds a trigger for nuclear war; none holds a kill switch. 

The risk of a single point of failure for planetary-scale destruction is tractable.  Leading experts have been repeatedly surprised by the effects of nuclear weapons, which suggests that careful study can further increase human knowledge.  Some Manhattan Project scientists suspected a nuclear weapon could ignite the Earth’s atmosphere.  The electromagnetic pulse caused by nuclear weapons was discovered through nuclear testing. The 1954 Castle Bravo nuclear test detonated with the explosive force of 15 million tons of TNT – 2.5 times what the designers expected.  The bioaccumulation of radioactive material (strontium-90) from nuclear testing in children was discovered by the St. Louis Baby Tooth Survey of 320,000 teeth donated from across the United States.  The idea of nuclear winter emerged from a 1983 study of the “Global Atmospheric Consequences of Nuclear War” by academic scientists.  Each of these important discoveries increased human understanding of the effects of nuclear war, but what we know is still dwarfed by what we don’t know.  The notional chart below suggests the scope and timescale of nuclear war effects that might relate to human extinction and the state of research progress toward this knowledge.  Information relevant to nuclear warfighting is relatively well understood, but almost nothing is known about secondary and tertiary effects of nuclear war over more than a decade.  Even modest research attention to these topics, on the order of a few million dollars, is almost certain to reveal important information we do not have about the converging risks of nuclear war to human survival over time.

3.  Distributed capacity for city-scale destruction

The cumulative potential of nuclear weapons to widely distribute capacity for city-scale destruction is existentially important.  The current nuclear weapons oligopoly of nine states won’t last.  The number of nuclear-armed states has increased over time; the forbearance of additional governments from acquiring nuclear weapons is similarly fragile.  The collapse of the former Soviet Union spread the world’s largest nuclear arsenal across several new international borders and foregrounded the risk of “loose nukes” (that must always accompany the existence of nuclear weapons).  Over enough time, other nuclear armed governments will fail.  No system for controlling nuclear weapons is perfect and terrorists have sought their own nuclear weapons.  In the long-term, nuclear weapons may distribute mass destruction piecemeal, one city or megadeath at a time.  H.G. Wells coined the term “atomic bomb” in his 1914 novel The World Set Free. Understanding that the destruction atomic bombs would make possible could not be limited to one place and time, he imagined a harrowing future in which anyone “could carry about in a handbag an amount of latent energy sufficient to wreck half a city.”  Wells predicted this immense destruction would happen over and over again at the hands of “bands of irreconcilables and invincible patriots, usurpers, adventurers, and political desperados…everywhere in possession of the simple apparatus,” and that would lead to an age of sickness, filth, starvation, and “everybody doing uncongenial things;” followed by a “pill-carrying age” to numb unbearable horror.  The occasional nuclear destruction of one or ten cities does not threaten human extinction, but the cumulative destruction of hundreds or thousands of cities over decades or centuries could have unknown cumulative effects like those of planetary scale nuclear war.

The potential to widely distribute capacity for city-scale destruction is neglected.  Wells rescues his fictional human race through an unspecified transformation in its ability to control nuclear technology that our real human race is making no effort to pursue.  Today’s nuclear armed governments plan investments ranging into the trillions of dollars for large numbers of diverse nuclear weapons systems. By comparison, research to detect, monitor, and verify the location or absence of nuclear weapons is much smaller. A National Academy of Science report last year on  “Nuclear Proliferation and Arms Control Monitoring, Detection and Verification: A National Security Priority” emphasized the need for such a research agenda, but actual investment in nuclear security is orders of magnitude lower than investment in nuclear arsenals.

The potential to widely distribute capacity for city-scale destruction is tractable.  A wide range of technologies that could contribute to human capability to ensure positive control over nuclear weapons, materials, and facilities is already emerging from the commercial sector.  Sensors, robotics, artificial intelligence, and trustless networks each provide crucial capabilities that could be applied to this problem.  A technology needs assessment convening the innovators developing these new technologies could chart a research agenda for their application to preventing distributed access to nuclear weapons; such a meeting might be accomplished for around $500,000.

4.  Changed geology permanently facilitating human access to nuclear weapons

Anthropogenic geological change making weapons-usable nuclear materials available is existentially important.  In the last eighty years, we have changed the composition of the Earth permanently making nuclear weapons much easier to build.  Acquiring enough separated plutonium or highly enriched uranium is the most significant technical barrier to nuclear weapons production.  Prior to the Manhattan Project, these materials simply didn’t exist in relevant quantities.  In 2021, the International Panel on Fissile Materials estimated the global stockpiles of highly enriched uranium and separated plutonium to be 1,255 and 545 metric tons respectively, while the Congressional Research Service reports that under 30 pounds of 90% enriched uranium or around 6 kilograms of plutonium would enough for an atomic bomb. Over the last eight decades, large quantities of these materials have been created, divided, subjected to multiple physical, chemical, and nuclear processes, built into nuclear weapons, and in some cases recovered from dismantled nuclear weapons and are now distributed in many locations, with tons characterized as “material unaccounted for.” Once produced, plutonium has a half-life of 24,000 years.  If no one ever separates another gram of plutonium from spent nuclear fuel, there will be more than enough remaining for a global thermonuclear war in the year 50,022 CE.  Even if a nuclear war or another catastrophe destroyed our entire civilization, successor civilizations could renew their pursuit of nuclear weapons with the knowledge that these weapons are possible and that large quantities of the necessary material have been left behind by our civilization.

Anthropogenic geological change having made weapons-usable nuclear materials available is neglected. Political attention and investment in nuclear material security have waned significantly since former President Barack Obama’s series of Nuclear Security Summits concluded in 2016.  Deep geologic repositories for nuclear waste are still in development, but not focused on the specific problem of weapons-usable nuclear material.  Protection of these materials still relies substantially on the security performance of individual governments.  Each nuclear armed government assumes it will last forever, and therefore does not consider the disposition of this material urgent or, in some cases, even appropriate (because it might be needed for nuclear weapons in the future).

Anthropogenic geological change having made weapons-usable nuclear materials available is tractable. Plutonium and highly enriched uranium can be rendered much less suitable for nuclear weapons.  Plutonium can be mixed with other materials and physically transformed to make its separation more difficult; highly enriched uranium can be blended down, returning it to its pre-Manhattan Project condition.  Advances in nuclear archeology technologies and methods hold great promise for working backward to ensure every bit of weapons-usable nuclear material is accounted and secured.  Investment on the order of a few tens of millions of dollars would revolutionize this field.

Nuclear war is not a one-time hazard.  It will be a persistent, global, distributed, and geologically entrenched drag on humanity’s prospects for survival and flourishing until humans resolve it.  Governments manage the dangers of nuclear weapons for their particular citizens on time horizons of years or decades, but the dangers posed by nuclear weapons to human survival are not limited to these brief times scales.  Investments on the order of a few millions of dollars in planetary-scale, longer-term vision can contribute substantially to mitigating these risks.