Nuclear war has returned to the realm of dinner table conversation, weighing on the minds of the public more than it has in a generation.
It’s not just “Oppenheimer’s” big haul at the box office: Since Russia’s invasion of Ukraine, the country’s officials have made nuclear threats. Russia has also suspended its participation in a nuclear arms control treaty with the United States. North Korea has launched demonstrative missiles. The United States, which is modernizing its nuclear weapons, shot down a surveillance balloon from China, which is building up its atomic arsenal.
“The threat of nuclear use today, I believe, is as high as it has ever been in the nuclear age,” said Joan Rohlfing, president and chief operating officer of the Nuclear Threat Initiative, an influential nonprofit group in Washington, D.C.
In this environment, a conventional crisis runs a significant risk of turning nuclear. It only requires a world leader to decide to launch a nuclear attack. And that decision making process must be better understood.
Historically, scholarship on nuclear decision making grew out of economic theory, where analysts have often irrationally assumed that a “rational actor” is making decisions.
“We all know that humans make mistakes,” Ms. Rohlfing said. “We don’t always have good judgment. We behave differently under stress. And there are so many examples of human failures over the course of history. Why do we think it’s going to be any different with nuclear?”
But growing scientific understanding of the human brain hasn’t necessarily translated into adjustments in nuclear launch protocols.
Now there’s a push to change that. The organization led by Ms. Rohlfing, for instance, is working on a project to apply insights from cognitive science and neuroscience to nuclear strategy and protocols — so leaders won’t bumble into atomic Armageddon.
But finding truly innovative, scientifically backed ideas to prevent an accidental or unnecessary nuclear attack is easier said than done. So is the task of presenting the work with adequate nuance.
Experts also need to persuade policymakers to apply research-based insights to real-world nuclear practice.
“The boundaries of that discourse are extraordinarily well protected,” said Anne I. Harrington, a nuclear scholar at Cardiff University in Wales, referring to internal pushback she says government insiders have faced when challenging the nuclear status quo. “So anyone who thinks that they’re going to make changes from the outside alone — I think that won’t happen.”
The world’s nuclear powers have different protocols for making the grave decision to use nuclear weapons. In the United States, absent an unlikely change to the balance of power among the branches of government, the decision rests with just one person.
“The most devastating weapons in the U.S. military arsenal can be ordered into use by only the president,” said Reja Younis of the Center for Strategic and International Studies in Washington, D.C., who is also a Ph.D. candidate in international relations at the Johns Hopkins School of Advanced International Studies.
In a crisis involving nuclear arms, Ms. Younis said, the president would probably meet with the secretary of defense, military leaders and other aides. Together, they would evaluate intelligence and discuss strategy, and the advisers would present the president with possible actions.
“Which could range from ‘let’s do nothing and see what happens’ to ‘let’s full-scale nuclear attack,’” said Alex Wellerstein, a professor at the Stevens Institute of Technology in New Jersey and head of a research project called “The President and the Bomb.”
In the end, though, only the president makes the call — and they can forgo guidance from advisers. A president could just press the proverbial button.
“These are the president’s weapons,” Ms. Rohlfing said.
Ahead of his electoral victory in 2016, experts and political opponents began raising concerns about investing in Donald J. Trump the power to order a nuclear attack. That debate continued in Congress through his term. By the time he left office, the Speaker of the House, Nancy Pelosi, openly asked the chairman of the Joint Chiefs of Staff to limit his ability to launch nuclear weapons.
It was in this milieu that Deborah G. Rosenblum, the executive vice president of the Nuclear Threat Initiative, invited Moran Cerf, a neuroscientist who is currently a professor at the Columbia Business School, to give a lecture to the organization in 2018. He titled it “Your Brain on Catastrophic Risk.” (Today, Ms. Rosenblum serves in the Biden administration as assistant secretary of defense for nuclear, chemical and biological defense programs — an office that briefs the president on nuclear matters.)
In a black T-shirt and jeans, Dr. Cerf briefed a room of experts and researchers on what brain science had to say about existentially troubling topics like nuclear war. The visit preceded a collaboration involving Dr. Cerf and a nonprofit called PopTech, whose conference Dr. Cerf hosts.
The groups, with a grant from the Carnegie Corporation of New York, are working to provide the government with science-based suggestions to improve nuclear launch protocols. Changing those policies is not impossible, but would require specific the right political scenario.
“You would need to have some sort of consensus that’s going to come from not just outside groups, but also policy and military insiders,” Dr. Harrington said. She added, “You probably also need the right president, honestly.”
The project includes a more public-facing arm: Dr. Cerf has been interviewing influential security experts like Leon E. Panetta, former secretary of defense and director of the C.I.A., and Michael S. Rogers, former director of the National Security Agency. Excerpts from these interviews will be cut into a documentary series, “Mutually Assured Destruction.”
With this project, Dr. Cerf and colleagues may have a conduit to share their findings and proposals with prominent government officials past and present. And he is optimistic about the difference those findings might make.
“I always think things will be better,” he said. “I always think that, with a nice smile, you can get the hardest opposition to listen to you.”
Dr. Cerf has the rapid cadence of a TED Talk speaker. Born in France and raised in Israel, he went to college for physics, got a master’s in philosophy, joined a lab that studied consciousness at Caltech and then transitioned to and completed a Ph.D. there in neuroscience.
Dr. Cerf said his primary critique of the system for starting a nuclear war is that despite advances in our understanding of the fickle brain, the status quo assumes largely rational actors. In reality, he says, the fate of millions rests on individual psychology.
One of Dr. Cerf’s suggestions is to scan presidents’ brains and gain an understanding of the neuro-particulars of presidential decision making. Maybe one commander in chief functions better in the morning, another in the evening; one is better hungry, the other better sated.
Other ideas for improving the protocols that Dr. Cerf has spoken about publicly generally can be traced back to existing research on decision making or nuclear issues.
Dr. Cerf says one important factor is speaking order during the big meeting. If, for instance, the president begins with an opinion, others — necessarily lower in the chain of command — are less likely to contradict it.
The idea that the hierarchical order of speaking affects the outcome of a discussion is not new. “That’s a classic experiment done in the ’50s,” said David J. Weiss, a professor emeritus at California State University, Los Angeles, referring to studies conducted by the psychologist Solomon Asch.
Dr. Cerf has also proposed decreasing the time pressure of a nuclear decision. The perception of a strict ticking clock to respond to a nuclear attack originated before the United States developed a more robust nuclear arsenal that might survive a first strike.
“We know that compressed time is bad for most decisions and most people,” Dr. Cerf said — an idea that goes back to at least the 1980s. Ideally, he says, if the United States received information indicating a launch, then the president could assess it and make a decision outside the direct heat of right-away.
The group’s main recommendation, though, mirrors proposals by other advocates: Require another person (or people) to say yes to a nuclear strike. Dr. Wellerstein, who did not contribute to the group’s research, says that such a person needs the explicit power to say no.
“Our belief is that the system we have, which relies on a single decision maker, who may or may not be equipped to make this decision, is a fragile and very risky system,” Ms. Rohlfing said.
While Dr. Cerf and colleagues have other papers in the works, the research from the project that he has produced doesn’t address nuclear weapons head-on. In one paper, participants made riskier decisions when they pretended to be merchants seeking deals on unidentified fruits of unknown value.
Dr. Cerf says that research is relevant to scenarios of high risk and low probability — like starting nuclear war — which often have numerous sources of uncertainty. A nuclear decision maker might be unsure of whether a missile is really in the air, how high a nuke’s yield is, why the missile was launched or whether more missiles will follow.
Another of Dr. Cerf’s studies involves climate change. It found that when people were asked to stake money on climate outcomes, they would bet that global warming was happening, and they were more concerned about its impact, more supportive of action and more knowledgeable about relevant issues — even if they began as skeptics. “You basically change your own brain without anyone telling you anything,” Dr. Cerf said.
He thinks the results could be applied to nuclear scenarios because you could use bets to make people care about nuclear risk and support changes to policy. The findings could also be used to evaluate the thinking and prediction of aides who advise the president.
Some scholars of decision science don’t agree on such extrapolations.
“To go from there to giving advice on the fate of the world — I don’t think so,” said Baruch Fischhoff, a psychologist who studies decision making at Carnegie Mellon University.
Paul Slovic, a professor of psychology at the University of Oregon and president of the nonprofit Decision Research, said that no psychological inquiry can stop at the experiment.
“You have to go back and forth between the laboratory studies, which are very constrained and limited, and looking out the window,” he said.
Experts say it’s also important to avoid selling too good a story about behavioral science to policymakers and elected officials.
“It’s just really easy to sell them stuff if you have enough bravado,” Dr. Fischhoff said.
Any brain, even a commander in chief’s, has a difficult time with the large-scale empathy required to understand what launching a nuclear weapon means. “We can’t really perceive what it means to kill 30 million people,” Dr. Cerf said.
There is a longstanding psychological term for this: psychic numbing, coined by Robert Jay Lifton. Just because humans are intelligent enough to master destructive weapons “does not mean that we’re smart enough to manage them after they’re created,” said Dr. Slovic, whose research has extended the concept of psychic numbing.
Compounding this effect is the difficulty of paying appropriate attention to all important information. And that compounds with the tendency to make a decision based on one or a few prominent variables. “If we’re faced with choices that pose a conflict between security and saving distant foreign lives to which we’re numb because they’re just numbers, we go with security,” Dr. Slovic said.
Dr. Slovic has also researched factors that tend to make people — including presidents — more likely to favor a nuclear launch. In one experiment, for instance, he found that the more punitive domestic policies a person supported — like the death penalty — the more likely the person was to approve of using the bomb.
Other researchers, like Janice Stein, a political scientist at the University of Toronto, have looked into scenarios where military officers show a reluctance to pass information up the chain of command that may trigger a nuclear launch.
That actually happened in 1983, when Col. Stanislav Petrov’s command center near Moscow received data suggesting the United States had launched intercontinental ballistic missiles. Colonel Petrov thought it could be a false alarm and decided not to send the warning to his superiors. He was right. Because the colonel feared a nuclear war fought under false pretenses more than he feared not retaliating, a third world war did not begin.
In the past, Dr. Wellerstein says, nuclear launch plans have adapted to changing circumstances, philosophies and technologies. And presidents have changed the protocols because of fears that emerged in their historical moments: that the military would launch a nuke on its own, that the country would experience a nuclear Pearl Harbor or that an accident would occur.
Perhaps today’s fear is that individual psychology governs a world-altering choice. Given that, working to understand how brains might work in a nuclear crisis — and how they could work better — is worthwhile.
What comes after the science — how to change policy — is complicated, but not impossible. Nuclear protocols may have a sense of permanence, but they’re written in word processors, not stone.
“The current system that we have didn’t fall out of the sky fully formed,” Dr. Wellerstein said.