“I will make a hypothesis that authoritarian regimes who believe people are weaknesses… cannot be trusted, that they will naturally gravitate toward totally automated [warfighting] solutions,” Deputy Secretary of Defense Robert Work said at the Center for a New American Security’s Inaugural National Defense Forum. He was outlining the rationale for the United States’ Third Offset strategy, which serves as a response to current developments towards automated warfighting amongst the modern great powers.
Russia is a great example of this technological change of pace. Early 2015 saw Russia present the partially autonomous Armata T-14 in a military parade last spring, but Russian leaders are also flirting with the automation of nuclear deterrence. Recently, Russia openly considered sixth-generation unmanned strategic bombers and revealed plans for the autonomous, long-range Status-6 nuclear torpedo.
If Deputy Secretary Work’s hypothesis is true, can the world expect to see entire nuclear arsenals deployed by software as well? This development path and its corresponding logic could rationalize the deployment of unmanned ballistic nuclear missile submarines across the world and alter the nuclear strategy that the world currently understands. The United States and others need to evaluate the threat and craft policies accordingly.
It is worth considering how the United States’ competitors may rationalize converting their modern second strike forces – fleets of manned ballistic nuclear missile submarines (SSBNs) – into robotic hair-trigger weapons platforms given this new affinity for autonomy and the advance of intelligence-gathering platforms. Deputy Secretary Work noted that the United States “strongly” believes “that humans should be the only ones to decide when to use lethal force,” but he concluded that other countries may not have developed their own definitive stances on automated forces, including nuclear weapon states.
There are more states today with nuclear weapons than during the Cold War, and they exist in an infinitely more complex age of geopolitical anxiety. While the United States and Russia are developing fifth generation designs, younger nuclear countries such as , India, and Pakistan are developing their own native nuclear weapons and platforms for the first time. If any of these states take into account the advances in satellite-based anti-submarine warfare (ASW) technology in their design process, unmanned SSBNs could become the plausible, asymmetric response for maintaining future platform effectiveness for these newly minted nuclear powers.
Improved computing power and satellite proliferation allow states to observe submarines at shallow depths using lasers, infrared, and synthetic aperture radar with increasing ability. Keeping SSBNs submerged at depth for extended periods is the simplest countermeasure to satellite detection, but the crews’ needs require submarines to approach the surface every three months if not more frequently to maintain communication. Emergence exposes them over time. Given these challenges, the near indefinite submergence of a crewless platform could become the preferred alternative to developing experimental countermeasures (e.g. thermal coating). However, unlike some automated systems, this path comes at the steep cost of real-time control.
Natural hurdles impede modern radio communication with underwater platforms, offering the most powerful reasoning against the adoption of an unmanned SSBN design. Seawater disperses normal electromagnetic communications, so states developed one-way communication relays using very low (VLF) or extremely low frequency (ELF) transmissions. Neither can relay complex tasks or operate a platform at depth because of their low bandwidth, and they are limited to instructing submarines to approach the surface. A strategically placed acoustic phone network though could provide a limited alternative where a platform could check-in with its users. This is why submarines need to operate with some degree of independence, and it is also why the crewless SSBN is conceivably more difficult to deploy in policy than an unmanned bomber for some countries.
The breakdown in communication adds a plethora of moral and technical issues that need to be addressed in the deployment of unmanned SSBNs. First, user states will need to grapple with the ethics of placing nuclear warheads into the hands of autonomous software and reducing the margin of error for mitigating escalation established over the last half century. Second, they will need to determine the pragmatism of reliably deploying an unmanned SSBN, keeping it functional while isolated from human contact, and insulating it from potential tampering. There will certainly be countries that reject the idea because of these problems, but other states like Russia may accept the costs of this new form of ill-conceived deterrent.
Regardless of whether a majority of states will reject them outright, Russia is considering the deployment of unmanned strategic forces. The unmanned SSBN could then become a product of that development process among Russian military leaders and other authoritarian countries. The unmanned SSBN is a plausible countermeasure against the improving efficacy of ASW satellites, but its use will cost states direct control of their nuclear weapons and put that authority in the hands of software. The autonomous developmental path that Russia and others are looking to take could very well lead to this new, unfamiliar nuclear reality by shortening the escalation ladder and raising global uncertainty in both strike ability and even culpability. The United States and others must set policies regulating the development of automated nuclear platforms and articulating reactions to other countries that may pursue them before the threat becomes tangible.
Bryan Ericsson is a recent graduate of the Elliott School of International Affairs in Security Policy Studies. He has interned with the National Defense University, the Institute for the Analysis of Global Security, and the Center for a New American Security for a variety of programs.