ON 6 AUGUST 1945, a nuclear bomb was dropped on the Japanese port city of Hiroshima. Three days later, Nagasaki suffered the same fate. Three-quarters of a century on, the full human toll is still unclear. In Hiroshima alone, some 75,000 souls were obliterated instantly, with many more deaths in the following months and years.
These are the only times nuclear weapons have been used in war; debates about the rights and wrongs continue. As we remember those who died, we might also usefully cast a wider view: on what the bombings meant for humanity, for our relationship with technology and for our perception of what we now call existential risks, those that threaten to irrecoverably damage our potential or extinguish us as a species.
Doing so can inform our response to dangers we are confronted with today. Whether it is the covid-19 pandemic, climate change or the emergence of new technologies such as artificial general intelligence, we are faced with threats that are, in their own way, just as great as the nuclear bomb – but also subtly different. Hiroshima was the start of a long, continuing learning process of understanding them.
Advertisement
Humans have probably talked about the end of the world for as long as we have talked. It is a common part of mythology, giving a sense of structure to history: there was a beginning, we live in the middle and there will be an end. But existential risks were, by and large, not practical matters, except to a few millenarian cults.
With the development of science came various realisations. The past was far, far vaster than we knew. There had been a time before humanity. Humanity was a species among others – and species could go extinct. Cosmic disasters, from asteroid impacts to supernovae, were real. Eventually, the universe’s energy might run out, dooming it to heat death.
These insights made for good popular science, and inspired plenty of science fiction. But such thinking as there was on human extinction focused on natural threats, and these were seen as remote and unlikely. Lecturing at the Royal Institution in London in 1902, the writer H. G. Wells declared that “worlds may freeze and suns may perish”, but we need not worry about such distant calamities. In 1933, the biologist J. B. S. Haldane echoed Wells’s point, suggesting that the only real threat would be if our species developed weapons and technologies too far. Luckily, he concluded, it was “wildly unlikely” we would ever unlock the power of the atom.
The next year, physicist Leo Szilard conceived of the nuclear chain reaction. He patented a reactor the following year. By 1942, the first nuclear reactor had been constructed in Chicago as part of the Manhattan Project to build an atomic bomb. Soon after, a second sun rose over Hiroshima.
Hiroshima was the moment we first became fully aware that humanity could cause devastation possibly even on a global scale. As the hydrogen bomb was developed over the following decade, this destruction became not just a possibility, but also a part of geopolitical strategy. We realised that our survival as a species now rested in our own foolish hands.
In the years since, growing environmental concerns have reinforced this perspective: there could be an end of nature and habitability driven by human actions. Today, we recognise that the covid-19 pandemic was enabled and exacerbated by our impact on the environment.
In 1945, the end of the world shifted from being a myth or mere scientific possibility to something very real, even banal, that could be caused by bad policy-making, meaningless stupidity or plain mistake. The end of humanity need not be an end that made sense, but just the end of sense: a total nihilistic threat.
This transformed how we related to our own ingenuity as an intelligent species. Previously rosy views of the march of progress took a battering. It highlighted another chilling possibility, too. One reason it took so long for people to take existential risk seriously in the scientific age was the assumption that humanoid civilisation is omniprevalent throughout our galaxy. If we wiped ourselves out on Earth, another intelligent species would simply pick up where we left off.
But in 1950, Enrico Fermi, chief architect of the atomic age, asked the question “Where is everyone else?”, referring to those other civilisations throughout our galaxy. The search for extraterrestrial intelligence was revving up, yet despite early optimism, it was returning just deep, ominous silence.
This seemed like an omen for humanity’s own troubled cold war trajectory. Perhaps we hadn’t seen evidence of other civilisations because, before they can make themselves visible, they wipe themselves out in thermonuclear holocaust. Technology itself might put a time limit on technological civilisations.
Scientists were soon listing other plausible technological threats, and how best to anticipate and mitigate them. This has lately led to existential risk emerging as a research field in its own right. It attempts to understand different kinds of catastrophic and existential risks, and then find their commonalities and effective ways to reduce them. We are just two of many people exploring these questions.
There is much still to learn. In the past, there was an unfortunate tendency to ignore or forget that which had been learned. An early instance occurred during the Manhattan Project. The physicist Edward Teller, a driving force behind the later creation of the hydrogen bomb, had warned that nuclear explosions could trigger fusion in the atmosphere, igniting it. Other researchers were sceptical, but serious debate and calculation convinced them the bomb was “safe” in that regard.
After the second world war, a report was published on the risk of atmospheric ignition, the first time a possible anthropogenic disaster was scientifically analysed, using several independent arguments to minimise the risk of the overall conclusion being flawed. It stated that “the complexity of the argument and the absence of satisfactory experimental foundations makes further work on the subject highly desirable”. The world was too busy with other things to take note: the hydrogen bomb’s development continued.
“Technology itself might put a time limit on technological civilisations”
What have we learned over the intervening decades? One central insight is that existential risks come in different forms, each posing different challenges. Nuclear weapons fall under the category of actual technological innovations. Other examples today are bioweapons or AI with human-like, or greater, capability. Here, the goal is to predict what future technology may be able to do and steer away from dangerous possibilities. As the cold war showed, something being dangerous for everyone doesn’t make it easy to coordinate against.
Other risks start out as disturbing scientific possibilities, like Teller’s concern about igniting the atmosphere. The fear that CERN’s Large Hadron Collider could make planet-eating black holes is a recent example. Whether there actually is a risk requires careful further analysis. This has been carried out for catastrophic asteroid impacts, where we now know the risk over the next century is manageably small, thanks to astronomical surveys. With that has come a recognition that it may be good to have some readiness for when a surprise impactor is discovered.
Still other risks are discovered –and their disaster potential recognised – only later. Examples include supervolcanoes, ozone depletion or dramatic climate change. Anthropogenic climate change was first predicted by Svante Arrhenius in 1896. Being a Swede, he mostly saw it as a good protection against a future glacial period. It wasn’t until the 1970s that we realised the actual danger.
Finally, there is a messy category of systemic risks. A huge solar flare might not kill anybody on Earth directly, but it could wreck electricity supplies for months. Much of the world is dependent on intricate online services. Food security requires perfect logistics, low fuel prices and a predictable climate. Things could go existentially wrong via a chain reaction of mishaps, disasters and conflicts. This is the hardest category to study and do something about.
And what of covid-19? Where does it fit into this taxonomy of risk? On its own, a naturally emerging pathogen such as the new coronavirus seems unlikely to cause the outright extinction of humanity; evolution provides a trade-off between lethality and transmissibility with natural diseases, if not with engineered ones. We have relatively few examples of infectious diseases wiping out or threatening to wipe out entire animal species. The current ravages of chytrid fungi among amphibian populations are an exception, but we know of almost no examples in mammals.
Instead, a pandemic’s true risk may be that it weakens society and compounds the likelihood of other threats being fatal. Certainly, our response to covid-19 proves how unprepared we still are for genuine large-scale perils.
It can take surprisingly long to recognise problems. Researchers have proclaimed artificial general intelligence to be around the corner since the 1950s, but almost nobody seems to have taken the risks seriously until the late 1990s. Before then, the only risks raised were some safety concerns with industrial robots. This is peculiar, since even human-level AI could be used for nefarious purposes, and there is no reason to think human intelligence is a ceiling. Perhaps researchers were privately more realistic about AI’s progress than their public pronouncements, but it is equally likely that they didn’t want to sound too outrageous or science fiction-like. We may be downplaying many other serious risks because we don’t want to rock the boat or because they sound too silly.
At other times, problems may be missed by accident. Ozone depletion from CFCs was predicted in the 1970s. However, because software screened out the extremely low measurements of ozone over Antarctica as errors, the hole in the ozone layer wasn’t spotted until 1985. The shocking new data helped spur political action. In 1987, the Montreal protocol phasing out CFCs was signed, a case where a global risk has been somewhat successfully handled. It can happen.
Some mid-20th century existential concerns have also receded, such as a systemic risk associated with widespread overpopulation and starvation. In 1968, the biologist Paul Ehrlich stated that, in the 1970s, “hundreds of millions of people will starve to death in spite of any crash programs embarked upon now”. It didn’t happen: the spread of contraception, education, better crop yields, trade globalisation and other factors rendered the prediction mostly wrong, although undernutrition and deprivation remain with us.
Indeed, concerns about underpopulation and population ageing are increasing. To some extent, the predictions about mass starvation were self-defeating because they motivated people to work out solutions, but these were largely just reasonable solutions that we would have wanted to pursue anyway.
“We may be downplaying serious problems because we don’t want to rock the boat or sound silly”
Even the threat of nuclear war receded a few times, through safety-increasing mechanisms such as non-proliferation treaties, the “red phone” between the Kremlin and the White House, disarmament and periods of detente. Sadly, it can also advance due to bad governance and inept diplomacy. We have been far more successful in feeding the world than getting rid of nuclear arms.
So how can we use all these insights? Decision-making has to be done under conditions of uncertainty, yet most of our institutions hate uncertainty and use it as an excuse for not reacting. The tough lessons of the covid-19 pandemic show the price of being paralysed by uncertainty. Yet planning for every conceivable disaster or blocking everything new as a precaution is also foolish.
Top-down solutions are needed for rapid, decisive responses, like a crash programme to deflect an asteroid, avert nuclear conflict or ensure alternative food sources to cope with an impending, decade-long “winter” caused by a nuclear firestorm or supervolcano eruption. They are also needed for strong unanimity, as with climate deals.
One lesson of covid-19 is that these solutions are best prepared long beforehand, rather than improvised on the spot. Creating and implementing emergency playbooks and building the capacity to coordinate responses internationally takes time. Often what is needed is top-down decision-making to promote solutions, rather than prescribing them from the start. This can include sponsoring research into technologies that defuse predicted risks from other technologies, such as bioweapons or AI, before they become powerful enough to be dangerous. Other examples are creating markets or international standards for solutions.
But solutions consist more of smaller building blocks than of heroic decisions, Hollywood movies notwithstanding. Finding and implementing useful solutions can also be bottom-up. The changes that helped the world avoid the 1970s overpopulation scenario are an example, as is the current solar power revolution, and the search for ways to shift the global economy to a low-carbon mode. These innovations don’t need a UN security council resolution supported by all. There might be other solutions that help us defuse the nuclear arsenals. The only way to know is to investigate.
None of this is a given: we must have the will to work on solutions to the existential threats we face. In that sense, perhaps the greatest cause for hope as we remember Hiroshima is that humanity still exists 75 years on.
