According to a new book by Labor frontbencher Andrew Leigh, humanity has a one-in-six chance of ending in the next hundred years. The biggest threats are not the ones he expected. By Mike Seccombe.
Andrew Leigh on humanity’s one-in-six chance of ending
What’s the worst that could happen? It’s one of those phrases that all too often presages disaster, like “What could go wrong?” It’s the declaration of the underprepared and overconfident.
“It’s a line used by the lackadaisical who are about to do something risky,” says Andrew Leigh, explaining the title of his new book.
“And I worried that we, as a species, might be too casually walking towards the cliff.”
Leigh, who was a professor of economics at the Australian National University before he quit to become the Labor member for the Canberra seat of Fraser in 2010, is among the most impressively credentialled members of federal parliament. He holds first-class honours degrees in arts and law from Sydney University, a master of public administration and a PhD in public policy from Harvard.
He is also almost certainly the most prolific member. His curriculum vitae, listing his books, journal articles, research grants, lecturing positions, academic appointments and other writings on economics, the law, technology, political science and more, runs to 17 closely spaced pages.
Even for such a polymath, though, his new book is an ambitious project. It sets out to assess the survival chances of the human race.
In successive chapters it explores four big threats: pandemic disease, climate change, nuclear warfare and runaway artificial intelligence. It collates the best available evidence, scientific data and expert assessments and considers the chances that one or other – or a combination of some of them – will destroy human civilisation.
There’s a one-in-six chance that we will not survive another 100 years, he writes.
“We are playing a game of Russian roulette with humanity’s future. Six chambers. One bullet.”
And then the book analyses a fifth risk, which cuts across and exacerbates the others – the global success of a brand of politics that advances short-term, simplistic solutions for complex problems; that is anti-intellectual, anti-institutional, anti-international and thrives on conflict. It goes by a rather benign descriptor: populism.
Hence the subtitle: Existential Risk and Extreme Politics.
Leigh’s What’s the Worst That Could Happen? is a sobering read, but also engrossing in its depth of data and anecdote. The Saturday Paper spoke with him about it this week.
Mike Seccombe First question: Why did you undertake such a big and frankly frightening project?
Andrew Leigh Well, I really like people. And I think it would be useful if we didn’t wipe out humanity. You know, we care about our kids and in some sense we care about our grandkids and great-grandkids. But then there’s potentially another 30 million generations of humans that could populate the planet before the sun blows up. The lives of pleasure and meaning that they could experience are just extraordinary. And yet there’s a one-in-six chance that humanity destroys itself in the next century, and that none of those 30 million generations gets to live. So that struck me as a pretty important problem. And I thought I would go through the key risks.
MS Your book explores four existential threats to humanity: pandemic disease, climate change, nuclear war and runaway artificial intelligence. The thing that surprised me most, given we’re talking about the odds, was that artificial intelligence had the highest probability of wiping us out. One in 10 in the next hundred years.
AL It surprised me too at the outset. But artificial intelligence is likely going to become better than humans’ some time in the next century. The experts differ on whether that’s 10 years or 100 years away, but no one really argues that it will be impossible for computers to outperform humans. Once that happens, there’s a real question as to whether the interests of that artificial intelligence will be aligned with ours or not.
And you don’t necessarily need a malign artificial intelligence. Professor Nick Bostrom [mathematician, physicist, philosopher and director of Oxford University’s Future of Humanity Institute] gives the example of an artificial intelligence that just wants to make as many paperclips as possible. And it crunches up our houses and cars, not because it dislikes us, but because they provide useful raw materials for making paperclips. So I think it’s really important that as we are building an artificial intelligence, we ensure that it has interests which accord with the sorts of values we care about.
MS Okay, so the greatest risk is that machines will do us in. But the other three big risks relate more directly to human activity.
AL There’s an element of chance in pandemics. So, the increasing spread of humans into animal environments is increasing the number of zoonotic diseases that jump the barrier from animals – pangolins and bats, for example – to humans. It’s very clear that Covid won’t be the last of those.
MS You suggested, though, that there was actually a much higher chance of an engineered disease posing a threat to humanity than a natural one. One in 30, compared with one in 10,000.
AL Yeah, absolutely. CRISPR–Cas9 [gene editing] has only been around less than a decade. And the ability of people in their laboratories to print new genes … that technology is fabulous in its potential but also creates the risk that terrorists might use it in order to weaponise diseases. Now, the real risk is something that is both contagious and deadly. Measles is very contagious but not very deadly. Ebola is very deadly but not very contagious. But if you got something which was both, then potentially you could have a Black Death scenario on your hands.
We should pause at this point to note that these estimates of the odds, adopted by Leigh, originated with another eminent academic, Toby Ord, a computer scientist and philosopher with the University of Oxford’s Future of Humanity Institute, whose work is focused on existential risk. According to Ord, the odds of nuclear war or climate change destroying human civilisation within the next hundred years are actually a great deal longer than those of weaponised germs or runaway artificial intelligence. He put them at one in 10,000. Still, Leigh was initially dubious about whether things were really so dire.
AL When I began the project I was honestly quite sceptical as to whether climate change could be a catastrophic risk. But I was persuaded by reading the work of the late Harvard professor Marty Weitzman [widely considered the world’s most influential environmental economist] that there is a very small chance of very big temperature increases. Maybe a 10 per cent chance of a 6 degree temperature rise, and 1 per cent chance of 10 degrees. And in that case, suddenly, the planet is radically transformed, and vast swaths of the planet become uninhabitable. I don’t think this is likely, but maybe it is enough of a possibility that it should significantly increase the case for strong climate action.
MS Well, exactly. Ten per cent is actually quite a significant chance. It is not a remote possibility.
AL Would you get on a plane with a 10 per cent chance of crashing or even a 1 per cent chance of crashing? No, of course not.
MS And nukes?
AL Across the world’s nine nuclear powers, there are around 14,000 nuclear weapons, about 13,000 held by the US and Russia. But it doesn’t take thousands of missiles to cause calamitous outcomes. A single nuclear weapon can devastate a city. One hundred nuclear weapons could kill more people than died in World War II.
A world with nine nuclear powers is considerably more volatile than a world with two powers. If MAD [mutually assured destruction, the theory saying that because no one could win a nuclear war, no one would have an incentive to start one] was problematic when applied to a pair of adversaries. It becomes even more fragile when several players are involved. When multiple people are pointing their guns at one another, one trigger pull can cause a chain reaction where everyone dies.
One point in Leigh’s book is that people – including many in government – don’t understand risk. People generally worry more about shark attacks than drowning in a bath. We’re more likely to fear dying in a plane crash than in an airport taxi. News reports often emphasise terrorism over family violence. In each case, those assessments are grossly wrong. One anecdote in particular underlines this misapprehension of risk. Leigh details the strong, bipartisan support within the United States congress for the expenditure of hundreds of millions of dollars to protect Earth from being hit by an asteroid. Between 2008 and 2021, the budget for “planetary defence” increased 30-fold. He quotes a question from Republican Senator Ted Cruz to NASA officials in 2018: “What steps do we need to be taking so that we don’t have to rely on sending Bruce Willis to space to save humanity?”
But the chances of an asteroid or comet wiping out humanity in the next century are one in a million. “If other existential risks received the same political priority as asteroid risk,” writes Leigh, “this book would be unnecessary.” Ted Cruz is deeply concerned about asteroids. But he denies the reality of climate change. He condemned the children’s television show Sesame Street because Big Bird advocated Covid-19 vaccinations. Cruz either doesn’t understand relative risk or chooses to ignore it for reasons of perceived political advantage.
MS Which brings us to the fifth risk: political failure. It was only 30 years ago that Francis Fukuyama declared “the end of history”. You know, that we’d reached the end point of ideological evolution and Western liberal democracy would be the final form of human government and would solve all our problems. But you detail a picture of democratic decline taking place across the world.
AL Well, there’s been pretty significant backsliding. We’ve seen populists elected everywhere from the Philippines to Hungary to Turkey to Poland. In 2019 and 2020, the world’s four largest democracies, with a combined population of over two billion, were run by populists: Trump in the United States, Modi in India, Widodo in Indonesia, and Bolsonaro in Brazil. And we’ve seen a decline in the share of countries that are rated as strongly democratic.
MS And you identify the ones who’ve got elected in recent years. These are mostly right-wing populists. But you get populists on both ends of the political spectrum, don’t you? Can you define right-wing as opposed to left-wing populism? Do they have basic tenets that are in some way different?
AL Populists regard politics as being a contest between a pure mass of people and a vile elite. For the left-wing populist, your classic Latin American 1970s populist, the vile elite is the super-rich. For right-wing populists, the enemy tends to be the intelligentsia – people with university degrees, urbanites, experts and immigrants. But it’s the demonisation that really matters. On both sides, populism is all about dividing rather than uniting.
MS You identified five factors that combined to make the 21st century the age of the populist: jobs, snobs, race, pace and luck. Do you want to expand a little bit on each of those categories?
AL In developed economies, middle-class jobs, those that pay enough to support a family and mortgage, have declined as manufacturing has declined and as many routine jobs have been automated out of existence. In the US, almost two-thirds of the jobs created since 1990 have been “low quality” in the sense that they paid less than the average wage. One study of European elections found that the higher the unemployment rate, the better far-right parties performed. Economic pain seems to benefit only the far-right.
By snobs I mean the way mainstream leaders have ignored the rise of populism. Think about the way in which the Republican establishment in the United States failed to see the rise of Trump or, indeed, the British Conservative establishment failed to foresee the rise of Boris Johnson.
In his book, Leigh cites another example of how establishment centrists have done a “lousy job of maintaining their supporter base” – the replacement of Malcolm Turnbull by Scott Morrison. The problem of snobbery, he writes, has led to mainstream progressives losing the support of less-educated voters. In the postwar era, parties of the left are increasingly likely to draw their support from highly educated voters.
Leigh notes that Trump once told a campaign rally, “I love the poorly educated.” In the postwar era, Leigh says, the US Democratic Party has moved from being the workers’ party to the party of the highly educated, and a “strikingly similar shift” has occurred in Britain, France, Italy, Switzerland, Canada, the Netherlands, Australia and New Zealand.
AL Race is the weaponising of diversity, pace the unsettling change in technology and society. And luck is simply the chance that when certain political entrepreneurs came along, they were cannier than their predecessors. There were people who preceded Donald Trump who had a lot in common ideologically, but Trump was just a more skilled political entrepreneur.
MS You are also critical of left-wing populists, like Bernie Sanders. But is it not fair enough, for example, given increasing economic inequality, to rail about the 1 per cent? I mean, on the data, there is a problem, why not rail against that? And why does it qualify as populism if you do?
AL Because when you attack institutions, and when you pretend to people that the entire system is corrupt and broken, it becomes much harder to build the public support to tackle long-term challenges. So if you want to address the challenge of nuclear conflict, you need a strong bureaucracy, you need international co-operation, and you need a level head and a long view. All of that becomes much harder if you turn up the temperature.
MS Okay, so the critique applies equally to both ends of the spectrum. We’re just seeing more success for the right. We see the rise of populism happening all over the world, and the United States serves as a particularly stark example. But it is not as pronounced in Australia. We have not seen the same degree of inequality here. We didn’t suffer greatly during the GFC. We didn’t suffer as much as many other countries during the Covid-19 crisis. Has our luck insulated us to some extent?
AL I don’t think Australia is immune from this. We’ve certainly seen a significant fall in the share of Australians voting for major parties and more people voting for populist alternatives. That gets bigger the further you go away from major cities. We’ve seen the use of race and racism to win votes in Australian elections. And we’ve seen, in the context of the anti-lockdown/anti-vaccination protests, a significant populist surge. So there’s significant reason to worry about Australia. And it may just be that the political launch or populist political entrepreneur, with the political skills of Trump, is yet to come around the corner…
So, to answer the question posed in the title of Leigh’s book, what’s the worst that could happen?
It is not actually climate change or rogue artificial intelligence. Nor is it pandemic or nuclear war, per se. It is the proliferation of political leaders who would rather pursue the politics of ignorance, fear and division for short-term advantage, rather than address the big, existential issues. That’s when it gets really bad. That, possibly, is how humanity ends.
This article was first published in the print edition of The Saturday Paper on Dec 18, 2021 as "The one-in-six chance".
A free press is one you pay for. In the short term, the economic fallout from coronavirus has taken about a third of our revenue. We will survive this crisis, but we need the support of readers. Now is the time to subscribe.
Letters & Editorial