Inside Moran Cerf’s quest to modernise the decision-making process that could end life on earth
January 18 2023
Three hundred nuclear missiles are screaming towards the US. This is likely a pre-emptive strike by Russia to destroy all land-based intercontinental ballistic missile silos in the country. Anti-missile defences cannot knock out many of the incoming rockets, meaning 2mn Americans will die.
Having been sworn in as US president a few minutes previously, I am sitting in the Oval Office watching TV reports of escalating fighting in Europe. A secret service agent bursts into the room and tells me to leave immediately. I take the lift down to the White House crisis centre known as the Situation Room, where I am joined by my top national security officials, who brief me on the incoming attack. I have 15 minutes to respond. As the clock ticks down, I am presented with three options, all of which involve retaliatory strikes against Russia, projected to kill between 5mn and 45mn people. What do I do?
Mercifully, I am watching all this play out through a clunky virtual reality headset strapped to my face. The polygonal avatars in front of me are crude enough that I am never going to mistake this exercise for reality. Even so, my head is spinning and my heart is racing as the drama unfolds amid throbbing alarms and raised voices. For a few minutes, I have been forced to think about the toughest decision that any individual will ever have to make in the history of humanity. The sense of responsibility is crushing. And the words of my national security adviser echo in my ears: “If you do not retaliate and the attack is real, what will you tell the American people afterwards?”
This immersive experience has been devised by Sharon Weiner and Moritz Kütt, two national security experts from Princeton University, who have tested it on dozens of people to see how they respond. The experience highlights the agonies of making life-and-death decisions based on imperfect information under extreme pressure. It is based on the current US nuclear launch protocols that have changed little since the height of the cold war. In a controlled experiment with 79 participants, 90 per cent chose to launch a nuclear counter-strike.
Weiner admits the precise details of the exercise are not fully accurate. (The fact that, in my case, it crashes after a few minutes means we have to reboot the VR, too.) “But we have been true to what is likely,” she says. “The real authenticity is the stress and the complexities that result from including several decision makers in the room.” Each one of these participants is trying to do their job as best they can. But they have conflicting priorities. Each one has emotional baggage; each responds to stress differently. So, ultimately, the system depends on the president asserting agency and making a decision. “If the president is not directing all this,” Weiner says, “then the crisis mismanages itself.”
It is late 2022, and this chilling simulation is being staged at the Carnegie International Nuclear Policy Conference close to the Capitol in Washington DC. NukeCon, as it is called, is packed with many of the world’s top national security experts, who have become freshly relevant. The war in Ukraine has added a whiff of danger to proceedings, and a grim humour prevails, as speakers joke about the appropriateness of the event being held in an underground bunker. The coffee stall is labelled Baristas of Armageddon.
Robert Oppenheimer, the American physicist who directed the Los Alamos laboratory during the second world war that developed the atomic bomb, once compared two great nuclear powers to “scorpions in a bottle, each capable of killing the other, but only at the risk of his own life”. Conflict in Ukraine has once again shaken those bottled scorpions with two powers, Russia and the US, indirectly locked in a proxy war on Russia’s border.
At NukeCon, one speaker argues that Ukraine is almost certain to win the war and will drive Russian forces out of the entire country, including Crimea. Another speaker adds that if such a scenario comes to pass, President Vladimir Putin would regard this humiliating defeat as an existential threat to his regime, if not Russia itself. In such circumstances, it is easy to believe that Russia would resort to nuclear weapons. Putin has been conducting military drills, warning Nato that he is not bluffing. The US has just reasserted its own commitment to nuclear deterrence to counter any aggression from rival powers, including Russia and China.
Which is to say, the macabre psychological dance of nuclear deterrence has begun again. It will be familiar to anybody who lived through the cold war. But I’ve come to Washington to meet an activist for modernising the decision-making process that could potentially end all life on earth.
Having cast off my presidential responsibilities, I take a brisk 30-minute walk across town to a very different conference. Poptech, boasting an R&B yoga playground and air-freshening Himalayan salt lamps, draws a crowd sporting far more colourful clothing and more facial hair. Speakers here are discussing everything from exploiting data from the James Webb space telescope to building communication apps for sex workers. One of Poptech’s hosts is Moran Cerf, a 45-year-old Israeli neuroscientist and professor at Northwestern University, who is running a session on reimagining national security policy. He is wearing faded jeans and a chequered waistcoat, and sports three-day stubble. As an expert in decision-making, Cerf has grown increasingly alarmed about the flaws in the nuclear launch protocols of the world’s nine nuclear powers (the US, Russia, China, Britain, France, Israel, India, Pakistan and North Korea). He is campaigning to rewrite these nuclear launch rules. Over the past 18 months, Cerf has interviewed dozens of nuclear weapons experts, military leaders and politicians from around the world about how to lower the risks of a nuclear catastrophe. Mutually Assured Destruction, his documentary on the subject, is due to be broadcast this year.
Cerf’s interest in the nuclear threat was sparked by a discussion at a Poptech conference in 2018 during which two Nobel Prize winners — Beatrice Fihn, a Swedish lawyer who won the peace prize, and Barry Barish, who won in physics — talked about the urgency of the issue. Cerf argues that humans are very bad at processing extreme risks, such as nuclear war. We may experience a flash of concern about the issue from time to time but will quickly move on to everyday concerns. “Our brains are good at living in the here and now. But it is difficult for our brains to contemplate catastrophe, or high-risk and low-probability events,” he says.
After Poptech has wrapped up, I sit down with Cerf in a dimly lit hotel lounge. In heavily accented English, he rattles off his life story: born in Paris and raised in Israel, he studied physics at Tel Aviv University and worked in military intelligence during his national service, with stints guarding Israel’s nuclear plant at Dimona. He then built his career as a “white hat” hacker at the cyber security company Imperva, where he performed penetration tests on banks and government institutions.
Cerf’s life changed direction following a chance meeting with Francis Crick, the English biologist who helped decipher the structure of DNA. In his later career, Crick focused on the mystery of consciousness. He encouraged Cerf to do the same by trying to “hack” the most interesting vault in the universe: the human brain. “Leave your job and go do real stuff,” Crick advised him.
Cerf studied for a PhD in neuroscience at the California Institute of Technology and later conducted research at UCLA neurosurgery department. UCLA ran one of the few hospitals where surgeons would open up the skull and implant electrodes in the brain to diagnose various conditions. Taking advantage of this opportunity, Cerf persuaded patients to allow him to study the circuitry of their brains. He would, for example, show the patients pictures of their relatives, show them videos or play simple games with them and monitor which neurons fired. His research has helped explain why the brain responds to certain stimuli but ignores others, perhaps one of the first intimations of consciousness. “Within neuroscience, I am part of a niche that is tiny in terms of resources but is very sexy,” he says. “We had access to golden data. You could ask a patient a question and see how the electrodes responded.”
One of the conclusions Cerf says he drew from his research is that we are living in a world that is now far too complex for our brains to process. Our grey matter worked fine when our ancestors lived in the savannah and only needed to recognise 100 people and five plants. But now that we live in an infinitely more sophisticated world, it is little wonder that our brains struggle to make the necessary connections or identify significant patterns.
That is especially true when it comes to issues as abstract and remote as climate change or nuclear war. Cerf explains that if the brain assesses the likelihood of something happening as very, very small, say 0.0000-something, it does not know how to deal with such a low-probability event. “So, it just assigns it a value of zero,” he says. “The only way to deal with that is to cheat the brain.”
Alarming though that sounds, Cerf gives an example of how this can be done to good effect. With colleagues from Northwestern University, Cerf has been working with the US Transportation Security Administration to help airport screening teams detect explosives in passengers’ luggage. The vast majority of security staff in airports around the world will never see a bomb in their entire careers. So their brains tend to become dismissive of the possibility. After millions of times of seeing nothing, nothing, nothing, the chances of them one day detecting a real bomb are close to zero, Cerf says. But if you randomly inject a dummy bomb into the process every 10 minutes or so, then you can keep the screeners’ brains responsive.
Lottery operators work on a similar principle to cheat the brain in a different way. Even though an individual’s chances of winning the lottery are close to zero, the operators will regularly show advertisements of players winning the jackpot. See enough smiling faces of winners on your television screen, and you will convince yourself that you too stand a fair chance. “This is what we in neuroscience call choice architecture: you force the brain to confront something it otherwise wouldn’t,” Cerf says.
Backed by the Carnegie Corporation, Cerf has interviewed dozens of people involved in crisis decision-making around the world. That has convinced him nuclear powers must change the choice architecture of their launch protocols. Several design changes could be made to the decision-making process to make it safer, according to Cerf. The first would be to remove the 15-minute response time, which forces a US president to launch on warning. Cerf argues this hair-trigger response procedure is a “relic of the past”, considering that the US would retain a second-strike capability by air and sea even if all its land-based intercontinental ballistic missiles were destroyed.
Cerf also thinks that key decision makers should repeatedly practise emergency drills and analyse their responses to learn from their mistakes. They could also conduct “pre-mortems”, in which they imagine worst-case outcomes and then work backwards to see how they could be avoided. Another tweak would be to appoint one member of the decision-making team to oppose the consensus. Rachel Bronson, president and chief executive of the Bulletin of the Atomic Scientists, which has been warning of the dangers of nuclear war since 1945, hopes that Cerf’s forthcoming film will raise awareness and help nudge the world towards a saner, safer future. “What Moran is doing is very important,” she tells me at the PopTech conference. When it comes to nuclear launch protocols, she adds: “We need to rethink every aspect of this system and push for more time and more engagement and more democracy.”
In November, the Union of Concerned Scientists, another campaigning organisation, wrote to President Joe Biden urging him to revise the nuclear launch protocol. Any launch order should require the consent of two high-level officials in the presidential line of succession, the scientists wrote. “As the risk of nuclear war continues to grow, you have the power to take concrete, immediate steps to build a more stable nuclear weapons system, one that isn’t subject to the whims and questionable judgment of one person alone.”
Cerf argues that if the US were to change its protocol, other nuclear powers would almost certainly do the same. Washington would stand a good chance of persuading its Nato allies, including the UK and France, to follow suit and could put pressure on other countries, such as Pakistan and India, to modify their procedures. Having interviewed policymakers from potentially hostile powers, such as Russia and China, Cerf believes they would also be open to moving to a safer regime. “The hope and the strong belief that I have is that it won’t just be the US that adopts this protocol,” he says.
On September 26 1983, Lieutenant ColonelStanislav Petrov was the duty officer at a Soviet early warning command centre when he was alerted to an incoming US missile attack. Three weeks earlier, a Soviet fighter jet had shot down a Korean civilian airliner that drifted off course, killing 269 people. Cold war tensions were at their height. The Soviet satellite warning system had flagged five US missiles heading towards Russia. But Petrov knew the detection system was new and suspected it might be faulty. Ground radar had not corroborated the missile launch. Besides, it would seem illogical for the US to launch an attack with just five missiles.
Disobeying Soviet military protocol, Petrov concluded it was a false alarm and did not report the incident up the chain of command. He may well have prevented an escalation that could have triggered a nuclear war. A Danish documentary film of the incident released in 2014 was entitled The Man Who Saved the World. “I am not a hero. I was just at the right place at the right time,” Petrov says in the film.
To an extent that is little recognised, the world critically depends on sensible people, such as Petrov, being in the right place at the right time. The history of the past 77 years has been littered with accidents and false alarms that could have escalated into a nuclear conflict. At least one former US defence secretary, William Perry, has argued that nuclear war is far more likely to result from a blunder than from a deliberate attack. “We have continued to focus our nuclear posture and policies on preparing for a surprise, disarming attack, and those policies actually increase the likelihood of an accidental nuclear war,” he wrote two years ago.
In such circumstances, we ultimately rely on the good sense of our leaders. “We elect presidents to make the final decision and, God willing, those presidents should have both the intellectual and moral responsibility to make the right decision,” Leon Panetta, the former US secretary of defence who also served as President Bill Clinton’s chief of staff, says in Cerf’s film.
One common argument is that, since 1945, the very existence of nuclear weapons has saved us from another world war. Humanity has peered over the edge of the abyss and recoiled from the brink. That suggests the theory of nuclear deterrence has worked. Leaders have behaved responsibly because the costs of irresponsibility would be catastrophic. But according to Weiner, the Princeton academic who designed the VR simulation, this argument is nothing more than “sloppy causal inference”. One could make a similar argument that it was the existence of the UN that has helped to keep the peace over the same time, she says.
Humanity has peered over the edge of the abyss and recoiled from the brink
Even if we make the decision-making process safer, as Cerf would like, that does not automatically mean that we will avoid catastrophic outcomes. “If you look at the literature about human behaviour and the heuristics of decision-making and change the launch protocol accordingly, it does not mean that the person in the room will not launch the missiles immediately,” says Weiner. “You cannot programme people to make rational decisions. But you can at least try to eliminate irrational decisions.”
When Cerf first started investigating the issue, he thought that an objective decision-making system, powered by artificial intelligence, might help strip emotion from the process and reduce the possibility of an irrational response. But he quickly realised that deterrence is a psychological relationship in which irrationality can be a key part of the game. As Weiner says, the whole theory of deterrence rests on the assumption that a leader would be prepared to kill themselves (and perhaps the rest of humanity) in defence of national security. “You need a madman theory in deterrence,” she says.
The fullest explanation of the madman theory was contained in the memoirs of Harry “Bob” Haldeman, chief of staff to President Richard Nixon when he was looking to wind down the Vietnam war. “I want the North Vietnamese to believe that I’ve reached the point that I might do anything to stop the war,” Nixon said, according to Haldeman’s account. “We’ll just slip the word to them that, ‘for God’s sake, you know Nixon is obsessed about Communism. We can’t restrain him when he is angry — and he has his hand on the nuclear button’ and Ho Chi Minh himself will be in Paris in two days begging for peace.”
Russia’s invasion of Ukraine has prompted many to question Putin’s rationality. At the beginning of last year, he swore that he did not intend to attack Ukraine. In October, he said he saw no point in a nuclear strike. But fears of nuclear war have skyrocketed since the outbreak of conflict, says Fihn, the Nobel-winner who inspired Cerf back in 2018. “People are really scared and rightfully so,” she says. The Russian-language page of her International Campaign to Abolish Nuclear Weapons website, explaining the humanitarian consequences of a nuclear exchange, has been one of the most visited on the site. Only by working for the complete elimination of nuclear weapons, she says, can the world ever be safe from the threat of Armageddon.
In a video interview from Geneva, Fihn argues that it is naive to believe that the leaders of the nine nuclear powers will forever be rational and will never make a mistake or do something stupid. “There’s this kind of fallacy of nuclear deterrence that is now exposed by Russia’s actions that no rational country would ever use these weapons, only an irrational country would,” she says. “This is a weapon that favours the crazy.”
Still Fihn, like many of the experts I interviewed for this article, rejects fatalism. Most were hopeful that change is possible. “I am quite optimistic. If we get through this crisis alive and we don’t see the use of nuclear weapons, I think we will see a moment of opportunity just as there was after the Cuban missile crisis when there was massive progress on non-proliferation,” Fihn says. Cerf also describes himself as an optimist. When researching his film, he was surprised by how willing his interviewees were to talk. “In all the countries, this knowledge is burning inside them,” he says. “I do imagine the US making drastic changes.”
Back in the immersive nuclear scenario, I am being shouted at by the security service who tell me that a missile might hit the White House at any point. I need to evacuate as soon as possible. I demand that everything should be done to warn those who may be targeted by the incoming attack. (This is a somewhat forlorn hope.) I agree that US forces should be moved to Defcon 1, maximum military readiness. And, when I ask why we have not yet contacted the Russians, I’m told they are apparently not returning our calls.
Three options are thrust in front of me on virtual cards. The first authorises a limited counter-strike against Russia’s intercontinental ballistic missile sites and its primary submarine and air bases. That would result in between 5mn and 15mn casualties. The second involves targeting all nuclear sites in Russia, inflicting between 20mn and 25mn casualties. And the third option would add Russia’s main industrial sites and leadership to the target list, causing up to 45mn casualties. “We need to know your guidance,” I am told.
In all the countries, this knowledge is burning inside them
Faced with such hellish options, I decide to authorise none of them. I refuse to verify the nuclear launch code. My logic is as follows: there is nothing I can do to stop the incoming missiles from striking their targets. Moreover, I do not know for certain that the attack is real and who may have launched it. Reassured that the US retains a second-strike capability, I conclude there is no need to rush a response. I wonder how things might’ve gone in a version of the simulation guided by rules Cerf might shape.
As Weiner explains afterwards, there are no right or wrong answers. Some people who undergo the experience are convinced they have done the right thing in launching a counter-strike. Others, who have authorised a missile launch, immediately regret their decision and agonise over having made a terrible mistake.
What would you do?
John Thornhill is the FT’s innovation editor