The Risks Posed by Emerging Technologies to Nuclear Deterrence
The global nuclear order is in a period of flux, arising in part from the latest information technology revolution. Enormous advances in computing power over the past few decades have facilitated the development of a wide range of new technologies that will increasingly affect the way that we think about nuclear weapons, nuclear strategy and nuclear deterrence. In particular, these so-called emerging technologies are transforming both the demand side of nuclear deterrence – that is, what type of threat needs to be deterred (in terms of both actors and particular types of capabilities), as well as the supply side – the type of weapons systems and capabilities that can be used to achieve this. These technologies are also creating a new ‘grey area’, or ‘grey zone’, between nuclear and conventional weapons that will serve to blur the deterrence equation still further. For much of the nuclear age, deterring a nuclear attack has been considered to be best achieved through the threat of nuclear retaliation. Today, however, this position appears to be shifting.
The concept of ‘emerging technologies’ is a loose – even nebulous – one, and is used widely by different people to mean slightly different things. For some, the emerging technology challenge is principally about a fairly narrow transformation in nuclear counterforce capabilities. For others, a broader definition might encompass so-called ‘killer robots’, quantum computing, rail guns, directed energy weapons, drones, 3D printing (which could become a proliferation issue), and perhaps even nanotechnology, which does not currently have an impact on nuclear deterrence but could do in the future. In order to provide some clarity when we think of the emerging technology challenge to nuclear deterrence, I suggest that we are essentially referring to two key developments:
- Ground-breaking advances in sensing technology (especially operating from space, but also under water) and the ability to process enormous caches of information are allowing for greater accuracy and tracking of military targets across different domains, as well as revolutionary advances in precision. The latter are driving more capable ballistic missile defences, and providing both conventional and nuclear weapons with greater sophistication, lethality and speed.
- The impact of artificial intelligence (AI), automation, and the myriad dynamics associated with the ‘cyber’ sphere are revolutionizing the way that nuclear and non-nuclear operations are and could be conducted, and, at the same time, the way in which nuclear systems could be held vulnerable to attack. These dynamics all potentially challenge the central importance of a secure second nuclear strike force and nuclear deterrence through the threat of nuclear punishment – the basis of nuclear deterrence thinking since the 1960s.
As regards the demand side of the nuclear deterrence equation, emerging technologies are broadening and reconstituting the types of threats (especially counter-force threats) that a state may need to deter. While deterrence of an attack using non-nuclear weapons has clearly always been part of nuclear thinking, developments in hypersonic missile technology, computer network operations (CNOs) and the increasing possibilities of autonomous weapons systems are all challenging this strategic picture. Hypersonic missile technology is – at its most basic – making the delivery of nuclear and conventional warheads quicker and more manoeuvrable, and theoretically far too quick and evasive to be intercepted by the current generation of ballistic missile defences. Thus, when nuclear-armed, hypersonic missiles can be seen as ensuring that an adversary’s weapons cannot be intercepted before they reach their targets, and (perhaps paradoxically) as reinforcing deterrence rather than undermining it. When conventionally armed – as noted below – they present a quite different challenge.
While ‘cyber’ is an awkward and nebulous descriptor, the threats posed by hackers breaking into critical infrastructure, including weapons systems, so that they fail to work, or at least fail to work as planned, is also a very different challenge to those encountered in the past. Whereas, previously, command and control nodes would probably need to be destroyed or damaged by kinetic weapons, similar objectives could potentially be achieved today by digital means, and possibly without the adversary knowing that their systems had been breached until it was too late. How to deter CNOs (a better descriptor than cyber) against nuclear weapons systems remains an ongoing debate, as does the challenge posed by third parties and non-state actors in the nuclear-digital space.
Advances in engineering and AI are creating the possibility of greater autonomy in the systems that must be deterred and defended against. The much-hyped ‘Poseidon/Status-6’ Russian underwater nuclear drone is a good example of this, but it is also possible to see greater autonomy across all nuclear systems in the future as technology allows.
On the supply side, emerging technologies are clearly reformulating the toolkit available to a state that can be used for deterring nuclear threats. The most obvious example here is ballistic missile defences. While the pursuit of active defence against nuclear attack is not a new thing, the ability to ‘hit a bullet with a bullet’ has been transformed in recent years by developments in both science and engineering. While not perfect, missile defences have increasingly become part of the deterrence and security planning of nuclear-armed states, and this clearly begins to cast doubt on the efficacy of waiting to strike second, especially for states with a limited nuclear-armed missile force. At the same time, processing power and sensing technologies have also driven the development of ever-greater precision for nuclear and non-nuclear weaponry, both regional and long range, as with the US Conventional Prompt Global Strike concept. These capabilities could theoretically be used to conduct non-nuclear strikes against critical nuclear targets, command and control apparatus, or the weapons systems themselves.
The same is also true for the possible use of CNOs against an adversary’s nuclear systems. The US has already begun a programme of ‘full spectrum missile defence’ or left-of-launch operations designed to target the missile and nuclear systems of potential adversaries before they can be used. The difference here, of course, is that it is much harder to assess or quantify the threat posed by intangible computer code than it is for a large and conspicuous nuclear-armed ballistic missile. A successful attack may only become known during a crisis.
It is much harder to assess or quantify the threat posed by intangible computer code than it is for a large and conspicuous nuclear-armed ballistic missile.
Finally, nuclear-armed and nuclear-powered submarines, long held as the bedrock of deterrence because they are so difficult to locate, are being challenged by underwater sensing technology that might potentially make them less invulnerable in the future. The same is also likely to be true when it comes to targeting mobile missiles. All of these are underpinned by developments in space capabilities and processing power, and – increasingly – developments in AI.
A transition in deterrence thinking?
When we take these developments together, it is entirely possible that we stand on the cusp of a transition in nuclear order and nuclear deterrence, and especially in deterrence through mutual nuclear vulnerability. The demand side of the equation is increasingly calling into question whether deterrence through the threat of nuclear punishment remains the best option available to deter existential threats. At the same time, various emerging technologies are transforming the supply side, or what is available beyond nuclear punishment in terms of a viable response. Both therefore suggest a possible return to a world where first strike becomes a common policy, potentially without having to employ nuclear weapons, and characterized as much by deterrence by denial (both defence and offence) as punishment. Of course, this will play out slightly differently in different nuclear relationships and regions, given differences in strategy, geography and emerging technological capabilities, but the trend is a general one that will impact how we conceptualize nuclear deterrence and nuclear order globally. This in turn will reshape how we conceptualize strategic stability, and place even more pressure on a deteriorating global arms control edifice.
We still live in a world where the threat of nuclear weapons and nuclear counter-value attacks are believed to deter, but the ways, ends and means – i.e. the methods, tactics, desired outcomes, resources and systems available for nuclear deterrence – are changing. Emerging technologies clearly present new challenges (both nuclear and non-nuclear) that must be met, and at the same time present a new suite of options for deterring nuclear attacks. While there have been earlier periods of significant technological change in the nuclear realm – with the development of silent submarines, stealth aircraft, cruise missiles, and so on – these all essentially served to reinforce the status quo. And the same is probably true for increases or innovation in nuclear capabilities such as hypersonics.
The challenge today, however, is qualitatively different, especially as concerns the integration and overlapping of new technologies. Many of these new capabilities are non-nuclear, and could augment or even replace nuclear weapons for certain deterrence functions, and many offer new counter-force options in the future. But perhaps most importantly, we are dealing with something that is not singular, but plural: emerging technologies rather than a particular technology. As a result, the impact is amplified considerably, and the challenge is therefore perhaps better presented as a shift in the context within which we think about nuclear deterrence. Consequently, we might need to rethink the central tenets of nuclear deterrence – and, particularly, what it is that we are seeking to deter, and how, in a very different world.
88 See for example, Lieber, K. and Press, D. (2017), ‘The New Era of Counterforce: Technological Change and the Future of Nuclear Deterrence’, International Security, 41(4), pp. 9–49, doi.org/10.7910/DVN/NKZJVT (accessed 14 Aug. 2019).
89 Kroenig, M. and Volpe, T. (2015), ‘3-D Printing the Bomb? The Nuclear Nonproliferation Challenge’, The Washington Quarterly, 38(3), pp. 7–19, doi:10.1080/0163660X.2015.1099022 (accessed 14 Aug. 2019).
90 The Economist (byline S. J.) (2019), ‘What are hypersonic weapons?’, 3 January 2019, https://www.economist.com/the-economist-explains/2019/01/03/what-are-hypersonic-weapons (accessed 14 Aug. 2019).
91 See Futter, A. (2018), ‘Cyber Semantics: Why We Should Retire the Latest Buzzword in Security Studies’, Journal of Cyber Policy, (3)2, pp. 201–16 (accessed 14 Aug. 2019).
92 Gady, F.-S. (2019), ‘Russia’s First ‘Poseidon’ Underwater Drone-Carrying Submarine to be Launched in 2019’, The Diplomat, 20 February 2019, https://thediplomat.com/2019/02/russias-first-poseidon-underwater-drone-carrying-submarine-to-be-launched-in-2019/ (accessed 14 Aug. 2019).
93 For example, see the US Nuclear Posture Review: Office of the Secretary of Defense (2018), Nuclear Posture Review.
94 Gormley, D. M. (2015), ‘US Advanced Conventional Systems and Conventional Prompt Global Strike Ambitions: Assessing the Risks, Benefits, and Arms Control Implications’, The Nonproliferation Review, 22(2), pp. 123–39, doi: 10.1080/10736700.2015.1117735 (accessed 14 Aug. 2019).
95 See Futter, A. (2016), ‘The Dangers of using Cyberattacks to Counter Nuclear Threats’, Arms Control Today, July/August 2016, https://www.armscontrol.org/print/7551 (accessed 14 Aug. 2019).
96 See Ingram, P. (2016), Will Trident still work in the future?, BASIC, https://www.basicint.org/wp-content/uploads/2018/06/Will-Trident-Work- Future-Jan2016.pdf (accessed 14 Aug. 2019).
97 Bracken, P. (2016), ‘Nuclear Stability and the Hunt for Mobile Missiles’, Foreign Policy Research Institute, https://www.fpri.org/article/2016/ 04/nuclear-stability-hunt-mobile-missiles/ (accessed 14 Aug. 2019).
98 Geist, E. and Lohn, A. (2018), How Might Artificial Intelligence Affect the Risk of Nuclear War?, Santa Monica, CA: RAND Corporation, https://www.rand.org/pubs/perspectives/PE296.html (accessed 14 Aug. 2019).