Pages

14 August 2022

The Beginning of History Surviving the Era of Catastrophic Risk

William MacAskill

We stand at the beginning of history. For every person alive today, ten have lived and died in the past. But if human beings survive as long as the average mammal species, then for every person alive today, a thousand people will live in the future. We are the ancients. On the scale of a typical human life, humanity today is barely an infant struggling to walk.

Although the future of our species may yet be long, it may instead be fleeting. Of the many developments that have occurred since this magazine’s first issue a century ago, the most profound is humanity’s ability to end itself. From climate change to nuclear war, engineered pandemics, uncontrolled artificial intelligence (AI), and other destructive technologies not yet foreseen, a worrying number of risks conspire to threaten the end of humanity.

Just over 30 years ago, as the Cold War came to an end, some thinkers saw the future unfurling in a far more placid way. The threat of apocalypse, so vivid in the Cold War imagination, had begun to recede. The end of communism a few decades after the defeat of fascism during World War II seemed to have settled the major ideological debates. Capitalism and democracy would spread inexorably. The political theorist Francis Fukuyama divided the world into “post-historical” and “historical” societies. War might persist in certain parts of the world in the shape of ethnic and sectarian conflicts, for instance. But large-scale wars would become a thing of the past as more and more countries joined the likes of France, Japan, and the United States on the other side of history. The future offered a narrow range of political possibilities, as it promised relative peace, prosperity, and ever-widening individual freedoms.

The prospect of a timeless future has given way to visions of no future at all. Ideology remains a fault line in geopolitics, market globalization is fragmenting, and great-power conflict has become increasingly likely. But the threats to the future are bigger still, with the possibility of the eradication of the human species. In the face of that potential oblivion, the range of political and policy debates is likely to be wider in the years ahead than it has been in decades. The great ideological disputes are far from settled. In truth, we are likely to encounter bigger questions and be forced to consider more radical proposals that reflect the challenges posed by the transformations and perils ahead. Our horizons must expand, not shrink.

Chief among those challenges is how humanity manages the dangers of its own genius. Advances in weaponry, biology, and computing could spell the end of the species, either through deliberate misuse or a large-scale accident. Societies face risks whose sheer scale could paralyze any concerted action. But governments can and must take meaningful steps today to ensure the survival of the species without forgoing the benefits of technological progress. Indeed, the world will need innovation to overcome several cataclysmic dangers it already faces—humanity needs to be able to generate and store clean energy, detect novel diseases when they can still be contained, and maintain peace between the great powers without relying on a delicate balance of nuclear-enabled mutually assured destruction.

Far from a safe resting place, the technological and institutional status quo is a precarious predicament from which societies need to escape. To lay the groundwork for this escape, governments must become more aware of the risks they face and develop a robust institutional apparatus for managing them. This includes embedding a concern for worst-case scenarios into relevant areas of policymaking and embracing an idea known as “differential technological development”—reining in work that would produce potentially dangerous outcomes, such as biological research that can be weaponized, while funding and otherwise accelerating those technologies that would help reduce risk, such as wastewater monitoring for pathogen detection.

The greatest shift needed is one of perspective. Fukuyama looked to the future a little mournfully, seeing a gray, undramatic expanse—a tableau for technocrats. “The end of history will be a very sad time,” he wrote in 1989, in which “daring, courage, imagination, and idealism, will be replaced by economic calculation, the endless solving of technical problems, environmental concerns, and the satisfaction of sophisticated consumer demands.” But at this beginning of history, this critical juncture in the human story, it will take daring and imagination to meet the various challenges ahead. Contrary to what Fukuyama foresaw, the political horizon has not narrowed to a sliver. Enormous economic, social, and political transformations remain possible—and necessary. If we act wisely, the coming century will be defined by the recognition of what we owe the future, and our grandchildren’s grandchildren will look back at us with gratitude and pride. If we mess up, they might never see the light of day.

THOSE WHO ARE YET TO COME

The fossil record indicates that the average mammal species lasts a million years. By this measure, we have about 700,000 years ahead of us. During this time, even if humanity remained earthbound at just one-tenth of the current world population, a staggering ten trillion people would be born in the future.

Moreover, our species is not the average mammal, and humans may well be able to outlast their relatives. If we survived until the expanding sun scorched the earth, humanity would persist for hundreds of millions of years. More time would separate us from our last descendants than from the earliest dinosaurs. And if one day we settled space—entirely conceivable on the scale of thousands of years—earth-originating intelligent life could continue until the last stars burned out in tens of trillions of years.

Far from being an idle exercise in juggling unfathomable numbers, appreciating the potential scale of humanity’s future is vital to understanding what is at stake. Actions today could affect whether and how trillions of our descendants might live—whether they will face poverty or abundance, war or peace, slavery or freedom—placing inordinate responsibility on the shoulders of the present. The profound consequences of such a shift in perspective are demonstrated by a striking experiment conducted in the small Japanese town of Yahaba. Before debating municipal policy, half the participants were asked to put on ceremonial robes and imagine they were from the future, representing the interests of the current citizens’ grandchildren. Not only did researchers observe a “stark contrast in deliberation styles and priorities between the groups,” the concern for future generations was infectious—among the measures on which consensus could be achieved, more than half were proposed by the imaginary grandchildren.

Thinking in the long term reveals how much societies can still achieve. As little as 500 years ago, it would have been inconceivable that one day incomes would double every few generations, that most people would live to see their grandchildren grow up, and that the world’s leading countries would be secular societies whose leaders are chosen in free elections. Countries that now seem so permanent to their citizens may not last more than a few centuries. None of the world’s various modes of social organization appeared in history fully formed. A short-term focus on days, months, or years obscures the potential for fundamental long-term change.

The fact that humanity is only in its infancy highlights what a tragedy its untimely death would be. There is so much life left to live, but in our youth, our attention flits quickly from one thing to the next, and we stumble around not realizing that some of our actions place us at serious risk. Our powers increase by the day, but our self-awareness and wisdom lag behind. Our story might end before it has truly begun.

HOW WE COULD END HISTORY

In contrast to Fukuyama’s “end of history,” other observers of international affairs have focused on the more literal meaning of the phrase: the potential for humanity to perish altogether. Such views were especially prevalent at the dawn of the Cold War, shortly after nuclear scientists enabled a massive leap in humanity’s destructive potential. As the British statesman Winston Churchill put it in 1946 with characteristic verve, “The Stone Age may return on the gleaming wings of science, and what might now shower immeasurable material blessings upon mankind, may even bring about its total destruction.” A few years later, U.S. President Dwight Eisenhower echoed these concerns during his first inaugural address, in which he warned that “science seems ready to confer upon us, as its final gift, the power to erase human life from this planet.”

Human history is rife with catastrophe, from the horrors of the Black Death to those of slavery and colonialism. But barring a few highly unlikely natural events, such as supervolcano eruptions or meteors crashing into the planet, there were no plausible mechanisms by which humanity as a whole could perish. In his book The Precipice, the Oxford philosopher Toby Ord estimated that even accepting all the most pessimistic assumptions, the accumulated risks of naturally occurring extinction still afford humanity an expected lifespan of at least 100,000 years.

Serious concerns about “existential catastrophe”—defined by Ord as the permanent destruction of humanity’s potential—emerged mainly in the second half of the twentieth century, hand in hand with an acceleration of technological progress. Lord Martin Rees, the former president of the Royal Society, wrote in 2003 that humanity’s odds of surviving this century are “no better than 50-50.” Ord estimated the likelihood of humanity wiping itself out or otherwise permanently derailing the course of civilization at one in six within the next hundred years. If either is right, the most likely way an American born today could die young is in a civilization-ending catastrophe.

Nuclear weapons exhibit several crucial properties that future technological threats may also possess. When invented in the middle of the twentieth century, they presented a sudden jump in destructive capabilities: the atomic bomb was thousands of times more powerful than pre-nuclear explosives; hydrogen bombs allowed for yields thousands of times again more explosive. Compared with the pace of increases in destructive power in the pre-nuclear age, 10,000 years of advances occurred within just a few decades.

These developments were hard to anticipate: the eminent physicist Ernest Rutherford dismissed the idea of atomic energy as “moonshine” as late as 1933, one year before Leo Szilard, another acclaimed physicist, patented the idea of a nuclear fission reactor. Once nuclear bombs had arrived, destruction could have been unleashed either deliberately, such as when U.S. generals advocated for a nuclear first strike on China during the 1958 Taiwan Strait crisis, or accidentally, as demonstrated by the harrowing track record of misfires in early warning systems. Even worse, measures to defend against a deliberate attack often came at the price of an increased risk of accidental nuclear Armageddon. Consider, for instance, the United States’ airborne alert, its launch-on-warning doctrine, or the Soviet “Dead Hand” system, which guaranteed that if Moscow suffered a nuclear attack, it would automatically launch an all-out nuclear retaliation. The end of the Cold War did not fundamentally change this deadly calculus, and nuclear powers still balance safety and force readiness at the heart of their policies. Future technologies might impose even more dangerous tradeoffs between safety and performance.

APOCALYPSE SOON?

But nuclear weapons are far from the only risks we face. Several future technologies could be more destructive, easier to obtain for a wider range of actors, pose more dual-use concerns, or require fewer missteps to trigger the extinction of our species—and hence be much harder to govern. A recent report by the U.S. National Intelligence Council identified runaway artificial intelligence, engineered pandemics, and nanotechnology weapons, in addition to nuclear war, as sources of existential risks—“threats that could damage life on a global scale” and “challenge our ability to imagine and comprehend their potential scope and scale.”

Take, for example, engineered pandemics. Progress in biotechnology has been extremely rapid, with key costs, such as for gene sequencing, falling ever faster. Further advances promise numerous benefits, such as gene therapies for as yet incurable diseases. But dual-use concerns loom large: some of the methods used in medical research could, in principle, be employed to identify or create pathogens that are more transmissible and lethal than anything in nature. This may be done as part of open scientific enterprises— in which scientists sometimes modify pathogens to learn how to combat them—or with less noble intentions in terrorist or state-run bioweapons programs. (Such programs are not a thing of the past: a 2021 U.S. State Department report concluded that both North Korea and Russia maintain an offensive bioweapons program.) Research published with pro-social intentions could also be misused by bad actors, perhaps in ways the original authors never considered.

The cranium and mandible of an early human species at the Museum of Human Evolution in Burgos, Spain, July 2010Felix Ordonez / Reuters

Unlike nuclear weapons, bacteria and viruses are self-replicating. As the COVID-19 pandemic tragically proved, once a new pathogen has infected a single human being, there may be no way to put the genie back in the bottle. And although just nine states have nuclear weapons—with Russia and the United States controlling more than 90 percent of all warheads—the world has thousands of biological laboratories. Of these, dozens—spread out over five continents—are licensed to experiment with the world’s most dangerous pathogens.

Worse, the safety track record of biological research is even more dismal than that of nuclear weapons. In 2007, foot-and-mouth disease, which spreads rapidly through livestock populations and can easily cause billions of dollars of economic damage, leaked not once but twice from the same British laboratory within weeks, even after government intervention. And lab leaks have already led to the loss of human life, such as when weaponized anthrax escaped from a plant connected to the Soviet bioweapons program in Sverdlovsk in 1979, killing dozens. Perhaps most worrying, genetic evidence suggests that the 1977 “Russian flu” pandemic may have originated in human experiments involving an influenza strain that had circulated in the 1950s. Around 700,000 people died.

Altogether, hundreds of accidental infections have occurred in U.S. labs alone—one per 250 person-years of laboratory work. Since there are dozens of high-security labs in the world, each of which employs dozens, perhaps even hundreds, of scientists and other staff, such a rate amounts to multiple accidental infections per year. Societies must significantly reduce this rate. If these facilities ever start tinkering with extinction-level pathogens, humanity’s premature end will be just a matter of time.

GOVERNANCE AT THE END OF THE WORLD

Despite this rising level of risk, it is far from assured that humanity will be able to take the necessary steps to protect itself. In fact, there are several obstacles to adequate risk mitigation.

The most fundamental issue is painfully familiar from the struggles of climate diplomacy in recent years. When burning fossil fuels, individual countries reap most of the benefits, but other countries and future generations will bear most of the costs. Similarly, engaging in risky biological research holds the promise of patentable drugs that could boost a country’s economy and prestige—but a pathogen accidentally released in that country would not respect borders. In the language of economists, imposing a risk on the future is a negative externality, and providing risk-reduction measures, such as establishing an early warning system for novel diseases, is a global public good. (Consider how the whole world would have benefited if COVID-19, like SARS between 2002 and 2004, had been contained in a small number of countries and then eradicated.) This is precisely the sort of good that neither the market nor the international system will provide by default because countries have powerful incentives to free-ride on the contributions of others.

Humanity has a number of avenues for escaping this structural tragedy. To assuage concerns about losing ground in the struggle for security, countries could enter into agreements to collectively refrain from developing especially dangerous technologies such as bioweapons. Alternatively, a coalition of the willing could band together to form what the economist William Nordhaus has called a “club.” Members of a club jointly help provide the global public good the club was formed to promote. At the same time, they commit to providing benefits to one another (such as economic growth or peace) while imposing costs (through measures such as tariffs) on nonmembers, thereby enticing them to join. For instance, clubs could be based on safety standards for artificial intelligence systems or on a moratorium on risky biological research.

Unfortunately, the resurgence of great-power competition casts doubt on the likelihood of these feats of global cooperation. Worse, geopolitical tensions could compel states to accept an increased level of risk to the world—and to themselves—if they perceive it as a gamble worth taking to further their security interests. (In the eight years during which the United States maintained bombers on continuous airborne alert, five aircraft crashed while carrying nuclear payloads.) And if even one state’s bioweapons program experimented with extinction-level pathogens—perhaps on a foolhardy quest to develop the ultimate deterrent—the next laboratory accident could precipitate a global pandemic much worse than that of COVID-19.

In the worst case, the great powers could, in their struggle for global hegemony, resort to outright war. For people who grew up in the West after World War II, this notion might seem far-fetched. The psychologist Steven Pinker has popularized the claim that violence—including among states—has long been on the decline. Subsequent analysis by the political scientist Bear Braumoeller and others, however, has substantially complicated the picture. The researchers have suggested that the intensity of conflict appears to follow what is known as a “power law,” meaning that after an interlude of relative peace, it is entirely possible that war might return in an even more deadly incarnation. Calculations by the computer scientist Aaron Clauset have indicated that the “long peace” that has followed World War II would need to endure for another century before it would constitute significant evidence of an actual long-term decline in war. Braumoeller asserted that it is “not at all unlikely that another war that would surpass the two World Wars in lethality will happen in your lifetime,” noting that in the conclusion of his book on the topic he “briefly considered typing, ‘We’re all going to die,’ and leaving it at that.”

Staving off the risk of World War III while also achieving unprecedented innovations in international governance is a tall order. But like it or not, that is the challenge we face.

INNOVATE TO SURVIVE

One response to this daunting challenge is retreat. If it is so difficult to safely govern emerging technologies, some argue, then why don’t we simply refrain from inventing them in the first place? Members of the “degrowth” movement take precisely this stance, decrying economic growth and technological progress as the main culprits behind alienation, environmental destruction, and all kinds of other harms. In 2019, 11,000 scientists from more than 150 countries signed an open letter demanding that the population of the world “be stabilized—and, ideally, gradually reduced” and that countries turn their priorities away “from GDP growth.”

Despite its intuitive appeal, this response is unrealistic and dangerous. It is unrealistic because it simply fails to engage with the interdependence of states in the international system. Even if the world’s countries came together temporarily to halt innovation, sooner or later someone would resume the pursuit of advanced technology.

Be that as it may, technological stagnation is not desirable anyway. To see why, note that new technologies can both exacerbate and reduce risk. Once a new technological danger has been introduced—such as by nuclear weapons—governments might require additional technologies to manage that risk. For example, the threat nuclear weapons pose to the survival of the human species would be greatly reduced if, during a potential nuclear winter, people were able to produce food without sunlight or if early warning systems could more reliably distinguish between intercontinental ballistic missiles and small scientific rockets. But if societies stop technological progress altogether, new technological threats may emerge that cannot be contained because the commensurate strides in defense have not been made. For instance, a wide variety of actors may be able to create unprecedentedly dangerous pathogens at a time when people have not made much progress in the early detection and eradication of novel diseases.

The status quo, in other words, is already heavily mined with potential catastrophes. And in the absence of defensive measures, threats from nature might eventually lead to human extinction as they have for many other species: to survive to their full potential, human beings will need to learn to perform such feats as deflecting asteroids and quickly fighting off new pandemics. They must avoid the fate of Icarus—but still fly.

The challenge is to continue reaping the fruits of technological advancement while protecting humanity against its downsides. Some experts refer to this as “differential technological development,” the idea being that if people can’t prevent destructive technology or accidents from happening in the first place, they can, with foresight and careful planning, at least attempt to develop beneficial and protective technologies first.

We’re already in a game of what Richard Danzig, the former U.S. secretary of the navy, has called “technology roulette.” No bullet has been fired yet, but that doesn’t change how risky the game is. There are many more turns to pull the trigger in the future: a bad accident and perhaps a fatal one is inevitable unless our species changes the game.

WHAT WE OWE THE FUTURE

Game-changers have so far been in short supply. Given the stakes, societies have to date done scandalously little to protect their future. Consider, for instance, the Biological Weapons Convention, which prohibits the development, storage, and acquisition of biological weapons. The national security expert Daniel Gerstein described it as “the most important arms control treaty of the twenty-first century,” yet it lacks a verification mechanism, and its budget is dwarfed by that of the Met Gala. As if this weren’t enough of a travesty, the BWC struggles to raise even the meager contributions it is due—a 2018 report by the convention’s chair lamented the “precarious and worsening state of the financial situation of the BWC . . . due to long-standing non-payment of assessed contributions by some States Parties.”

The management of nonbiological risks doesn’t inspire confidence, either. Research aimed at preventing the loss of control over artificially intelligent systems remains a minuscule fraction of overall AI research. And militaries are using lethal autonomous weapons on the battlefield, while efforts to limit such weapons systems have stalled for years at the UN. The domestic situation doesn’t look much better—less than one percent of the U.S. defense budget is dedicated to biodefense, and the majority of that goes to fending off chemical weapons such as anthrax. Even after COVID-19 killed one in every 500 people in the world and inflicted $16 trillion worth of economic damage in the United States alone, Congress couldn’t agree to provide a modest $15 billion to bolster pandemic preparedness.

This kind of risk reduction is so neglected that opportunities for positive change abound. One success story of existential risk mitigation is NASA’s Spaceguard program. At a cost of less than $5 million per year, between its inception in 1998 and 2010, scientists tracked more than 90 percent of extinction-threatening asteroids, in the process increasing the accuracy of their predictions and reducing the best estimate of the risk that one will strike the earth by a factor of ten. Consider also that during the COVID-19 pandemic, the U.S. government spent $18 billion on Operation Warp Speed to accelerate vaccine development. The program resulted in safe and effective vaccines that the United States and other countries were able to buy at a price constituting a small fraction of the vaccines’ social benefits, which have been estimated to amount to tens of trillions of dollars. The economist Robert Barro has estimated that between September 2021 and February 2022, these vaccines saved American lives at a cost of between $55,000 and $200,000 each, more than 20 times above the cost-effectiveness threshold that lifesaving policies usually need to meet.

A climate change demonstration in London, September 2019Hannah McKay / Reuters

If the world’s best and brightest step up and governments or the private sector provide funding, we can achieve even more impressive successes. For instance, although it still must overcome major technical hurdles, widespread metagenomic sequencing of wastewater would help detect novel diseases at a stage when they can still be contained and eradicated. The Nucleic Acid Observatory, based at the Massachusetts Institute of Technology, is pursuing just this vision. The public and private sectors should also develop better personal protective equipment and do further research on sterilization technology such as Far UVC—an ionizing radiation process that, if successful, could offer a near-universal defense against pathogens and be installed in any building. Regarding artificial intelligence, research aimed at making systems safe and reliable must be scaled up tenfold. The common thread running through measures such as these is an emphasis on defensive strategies that do not themselves create or enhance other risks.

Progress is also possible in other domains. Intelligence collection and analysis aimed at the known sources of large-scale risks will be critical. And although achieving complete certainty is impossible (as the astronomer Carl Sagan once quipped, “Theories that involve the end of the world are not amenable to experimental verification—or at least, not more than once”), scanning and forecasting what is on the horizon can help identify new concerns. In this vein, it is encouraging that the most recent Global Trends report by the National Intelligence Council included a discussion of the concept of existential risk, calling for “the development of resilient strategies to survive.”

More governments, institutions, and firms need to take such ideas seriously. Regulatory reform will also be important. In Averting Catastrophe, Cass Sunstein, a former head of the regulatory office at the White House, showed how the government’s current approach to cost-benefit analysis can’t sufficiently account for potential catastrophic risks. Sunstein argued for what he called the “maximin principle”: in the face of sufficiently extreme risks—and human extinction certainly qualifies as such—governments must focus on eliminating the very worst outcomes. As it happens, the White House is currently modernizing its framework for reviewing regulation. It should use this opportunity to make its approach to dealing with low-probability risks of extreme damage fit for the twenty-first century, whether by adopting Sunstein’s maximin principle or something similar that takes global catastrophic risks seriously.

Fukuyama prophesied “centuries of boredom at the end of history.” Nothing could be further from the case. Powerful and destructive technologies will present an unprecedented challenge to the current political system. Advanced AI could undermine the balance of power that exists between individuals and states: an entirely automated workforce would give the government little reason to treat its citizens well; a dictatorship that possessed an AI army and police force could prevent the possibility of an uprising or a coup. Government could use the prospect of a third world war as a reason to expand the state and crack down on individual liberties such as free speech on the grounds of protecting national security. The possibility of easily accessible bioweapons could be used to justify universal surveillance.

With humanity’s future in mind, we should resist such pressures. We must fight to ensure both that we have a future and that it is a future worth having. The cultural shift toward liberalism over the past three centuries created an engine of moral progress that led to the spread of democracy, the abolition of slavery, and expanded rights for women and people of color. That engine can’t be turned off now. If anything, we need to go much further in promoting moral and political diversity and experimentation. Looking back millennia, moderns see the Romans’ practices of slaveholding, torture for entertainment, and ultra-patriarchy as barbaric. Perhaps future generations will see many of our current practices as little better.

So we must walk a tightrope. We must ensure that global cooperation reduces the risks of global catastrophe to near zero while maintaining the freedom and diversity of thought and social structures that would enable us to build a future that our grandchildren’s grandchildren would thank us for. Contemplating large-scale political change is daunting, but past innovations in governance, such as the UN system and the EU, provide reasons for hope.

We are not used to seeing ourselves as one of history’s first generations; we tend to focus on what we have inherited from the past, not what we could bequeath to the future. This is a mistake. To tackle the task before us, we must reflect on where we stand in humanity’s full lineage. We in the present day recklessly gamble, not just with our lives and our children’s lives but with the very existence of all who are yet to come. Let us be the last generation to do so.

No comments:

Post a Comment