No, the U.S. Won’t Respond to A Cyberattack with Nukes

The idea that the U.S. is building new low-yield nuclear weapons to respond to a cyberattack is “not true,” military leaders told reporters in the runup to the Friday release of the new Nuclear Posture Review.

“The people who say we lowered the threshold for the use of nuclear weapons are saying, ‘but we want these low-yield nuclear weapons so that we can answer a cyberattack because we’re so bad at cybersecurity.’ That’s just fundamentally not true,” Gen. Paul Selva, vice chairman of the Joints Chiefs of Staff, said Tuesday at a meeting with reporters.

It’s an idea that military leaders have been pushing back against since the New York Times ran a Jan. 16 story headlined, “Pentagon Suggests Countering Devastating Cyberattacks With Nuclear Arms.”

When would the U.S. launch a nuclear attack in response to a non-nuclear event? The Defense Department says the threshold hasn’t changed since the Obama administration’s own nuclear posture review in 2010, but a draft of the new review that leaked online caused a bit of drama in its attempts to dispel “ambiguity.”

The new review gives examples of “non-nuclear strategic attacks,” Robert Soofer, deputy assistant secretary for nuclear and missile defense policy, told reporters on Thursday. “It could be catastrophic attacks against civilian populations, against infrastructure. It could be an attack using a non-nuclear weapon against our nuclear command-and-control [or] early-warning satellites. But we don’t talk about cyber.”

In his own conversation with reporters, Selva broadened “early warning” systems to include ones that provide “indications of warning that are important to our detection of an attack.” He also emphasized, “We never said ‘cyber.’”

There’s a reason for that. While cyberattacks on physical infrastructure can be very dangerous, they are unlikely to kill enough people to provoke a U.S. nuclear response.

An National Academies of Science and Engineering analysis of the vulnerability of U.S. infrastructure makes that point. A major cyberattack could cut off electrical power, resulting in “people dying from heat or cold exposure, etc.,” said Granger Morgan, co-director of the Carnegie Mellon Electricity Industry Center and one of the chairs of the report.  “A large outage of long duration could cover many states and last for weeks or longer. Whether and how many casualties there could be would depend on things like what the weather was during the outage.”

It’s a huge problem but not an event resulting in tens of thousands of immediate deaths.

Contrast that with a nuclear attack on a city like Moscow, even one using a device of 6 kilotons, much smaller than the ones the United States used against Japanese targets in World War II. The immediate result: there would be 40,000 deaths, according to the online nuclear simulation tool NukeMap.

Russia has demonstrated a willingness to take down power services with cyberattacks, as they did in Ukraine on Christmas Eve 2015. But these attacks were brief and occured in the context of actual fighting.

In other words, the worst cyber physical attack that top experts believe credible likely does not meet the threshold that the Defense Department has set out for deploying a nuclear weapon.

A somewhat more likely scenario: an attack that interferes with the ability of the U.S. to access data from the Space-Based Infrared System or other some other early warning indicator. By the Defense Department’s reasoning, that might draw a nuclear response. But that’s a specific attack that could portend a nuclear first strike.

As the U.S. adds more satellites, sensors, ground stations and other pieces of technology to its early warning network, the chances of a mishap rise — a troubling prospect that we’ve covered. But it’s a separate issue.

But isn’t the potential for a cyber-physical attack becoming greater? Asked about it on Thursday, Greg Weaver, the Joint Staff deputy director of strategic capabilities, said, “Not sure I agree with that.”

He added, “There’s no intent to expand the range of circumstances.”

Said Soofer, the deputy assistant defense secretary: “We maintain in this document, as in other documents, that we retain some level of ambiguity that we would use nuclear weapons. And so when you talk about using nuclear weapons to deter non nuclear attacks, it’s all going to be situationally dependent. It’s all contextual. If an attack that doesn’t cause a lot of casualties …and the President has a lot of tools with which to respond. … conventional tools, nuclear tools, cyber tools. He or she will make that decision based on the circumstances.”

On Friday, John Rood, defense undersecretary for policy, was asked whether a service disruption would prompt a nuclear retaliation. He did not directly respond. But he did say that the threshold for launching a nuclear strike is going up, not down. Before launching a strike in response to a cyberattack on a power plant, for instance, U.S. officials would want a lot of questions answered. “In the hypotheticals you cite, would that also involve the use of biological weapons against the U.S. population or allies? Would it involve the use of chemical weapons against our people? Would it involve a conventional attack against the U.S. or our allies in other parts of the world? The context in which an attack occurred would be how we would evaluate an appropriate response.”