December 2022 | By Natasha E. Bajema and John Gower
Overview
Emerging and disruptive technologies (EDTs) are expected to produce a variety of new risks for nuclear decision-making in the future, which need to be thoroughly explored, understood, tested, and then mitigated. In the next crisis between nuclear-armed states, decision-makers may confront a much denser fog of war, an increased pressure to act, and a grave lack of understanding about the features of their operating environment, to name a few of the changes on the horizon.
In this handbook, we explore the main technological effects of EDTs that may undermine nuclear decision-making in the future:
- the disruption of information flows;
- the illumination of behavior patterns;
- the compression of decision timeframes;
- the transformation of decision contexts.
Each of these technological effects may exert a negative impact on key assumptions of nuclear deterrence strategy and thus could fundamentally change the context for nuclear decision-making, increasing the risk of miscalculation, misinterpretation and escalation. We also examine how the effects of EDTs may exacerbate certain cognitive biases by providing or removing essential cognitive tools (information and knowledge), and explore how all of these factors might together impact nuclear decisions.
We conclude by providing an overview of risk reduction measures that span policy approaches such as technology regulation, technical tools such as using blockchain for authentication, and arms control and risk reduction measures designed to specifically reduce the risk of nuclear weapons.
Explaining the profound potential of EDTs for shaping the decision-making context and process is no easy task for any number of reasons. Social media, artificial intelligence, cyber weapons, sensors, drones, satellite imagery, hypersonic weapons, autonomous systems, and quantum computing all exist in various stages of development. Each technology will mature on separate timelines and likely start producing effects at different points in time. This makes it difficult to anticipate what combinations of EDTs may have an impact on decision-making during the next nuclear crisis.
Moreover, the divergent characteristics of EDTs suggest a significant variance in their impact on decision-making; some may simultaneously provide new opportunities for mitigating risks. In any future nuclear crisis, individual EDTs will not likely exert effects on their own. Rather, various combinations of EDTs may simultaneously produce consequences for decision-making, complicating efforts to train leaders on their new risks for nuclear weapons. Without an in-depth understanding among decision-makers of their effects, EDTs undermine prevailing concepts of nuclear deterrence and increase the risk of nuclear war.
At this early stage of understanding the challenge, the effects of EDTs on nuclear decision-making have not been sufficiently studied, largely due to the absence of tangible data and precedent. Much of the existing analysis to-date has thus been highly speculative and based on a limited number of historical case studies in which decision-makers considered the use of nuclear weapons, and for which records have been declassified.
Finally, while technological factors are important, they do not provide decision-makers with an exclusive basis upon which to formulate decisions to use nuclear weapons in a crisis. Many other factors are at play throughout such a decision-making process in the nuclear weapons domain, making it difficult to anticipate and isolate the actual effects of EDTs. As a result, it remains unclear how and when these new technologies might interact with nuclear weapons in a crisis.
To prevent such technological complexity from overwhelming nuclear decision-makers in the midst of the next crisis, we agree with Dr. Brad Roberts, former Deputy Assistant Secretary of Defense for Nuclear and Missile Defense Policy on the urgent need to prepare them now: “conceptually, organizationally, and operationally.” This represents a key objective of this project conducted by the Council on Strategic Risks (CSR) and the European Leadership Network (ELN) and serves as the main focus of this research.
In this handbook, CSR presents a comprehensive framework for thinking about technological complexity, nuclear decision-making, and risk reduction. This handbook intends to serve as an analytical guide for nuclear decision-makers to help them grapple with the technological complexity of the future, and further reduce the risk of nuclear war.
2.3.3.a Graphic Illustration
To summarize, we posit that EDTs in various combinations will exert certain types of effects through the lens of nuclear deterrence strategy. They will shape how a crisis unfolds through the different stages on the spectrum of conflict, alter features of the operating environment, and influence the decision-making process and mindset, thus having a potential impact on the outcomes of nuclear decisions. Since the mindset of selected decision-makers is non-trivial for nuclear decisions, we must examine how the effects of EDTs shape the decision-making context, undermine key assumptions of nuclear deterrence strategy, and exacerbate known cognitive biases.
The different layers of our framework include 1) relevant emerging technologies and their effects; 2) the mindset of decision-makers and features of the decision-making process; and 4) the nuclear decisions themselves. To help visualize the framework, we have developed a graphical representation. The illustration shows the proposed flow of impact of EDTs, their different types of effects on the decision-making context and mindset and their potential linkage to the outcomes of nuclear decisions.
Table of Contents
Chapter 1: Introduction
1.1. Overview 1
1.2. Lessons from the Past for the Future 4
1.2.1. Overhead Reconnaissance 5
1.2.2. Imagery Expertise 7
1.2.3. Information Sources 10
1.2.4. Communication Channels 12
1.2.5. The Geopolitical Context 15
1.3. Foundation for Curriculum Design 17
Chapter 2: A Framework for Technological Complexity
2.1. Overview 21
2.2. Literature Review 23
2.2.1. A Report from the NTI’s Nuclear-Cyber Weapons Study Group 24
2.2.2. RAND Workshops on Artificial Intelligence and Nuclear War 26
2.2.3. SIPRI’s Study on the Impact of AI on Strategic Stability and Nuclear Risk 27
2.2.4. CSIS Study on Situational Awareness and Crisis Decision-making 29
2.2.5. King’s College London Study on Social Media and Crisis Decision-making 31
2.2.6. King’s College London Study on Weapons of Mass Distortion 32
2.3. The Framework 34
2.3.1. The Fundamentals of Nuclear Deterrence Strategy 35
2.3.2. The Effects of Emerging and Disruptive Technologies 42
2.3.3. Decision-maker Mindset, Decision-making Process, and Decisions 45
2.4. Curriculum Design and Next Steps 49
Chapter 3: Emerging and Disruptive Technologies
3.1. Overview 51
3.2. Disrupting Information Flows 53
3.2.1. Social Media 54
3.2.2. Deep Fakes 62
3.2.3. Cyber Weapons 64
3.2.4. Electronic Warfare 69
3.2.5. The Impact of Disrupted Information Flows 70
3.3. Illuminating Behavior Patterns 71
3.3.1. Cyber Weapons 72
3.3.2. Social Media 73
3.3.3. Sensors 77
3.3.4. Drones (Unmanned Systems) 80
3.3.5. Satellites 82
3.3.6. Machine Learning 85
3.3.7. Impact of Illuminated Behavior Patterns 92
3.4. Compressing Decision Timeframes 93
3.4.1. Social Media 95
3.4.2. Machine Learning 96
3.4.3. Autonomous Systems 98
3.4.4. Hypersonic Weapons 101
3.4.5. Quantum Computing 103
3.4.6. Impact of Compressed Decision Timeframes 104
3.5. Transforming Decision Contexts 105
3.5.1. Navigating the Cyber-Nuclear Escalation Dilemma 107
3.5.2. Exploiting the Asymmetries of Artificial Intelligence 110
3.5.3. The Race for Machine Speed on the Battlefield 112
3.5.4. Impact of Transformed Decision Contexts 113
3.6. Impact on Nuclear Decision-making 114
3.7. Discussion Questions and Thought Exercises 131
Chapter 4: Nuclear Decision-making
4.1. Overview 137
4.2. The Decision-making Mindset and Process 139
4.2.1. Cognitive Biases 141
4.2.2. Too Much Information (Problem of Information Overload) 146
4.2.3. Not Enough Meaning (Problem of Uncertainty) 152
4.2.4. Need to Act Fast (Problem of Speed) 161
4.2.5. What Should We Remember (Problem of Flawed Memory) 166
4.3. Nuclear Decisions 168
4.3.1. Pre-Emptive Strategic Attack 171
4.3.2. Pre-Emptive Tactical / Local Attack 173
4.3.3. Reactive Tactical / Local Response 175
4.3.4. Reactive Strategic Response 176
4.3.5. Impact on Nuclear Decisions 177
4.4. Discussion Questions and Thought Exercises 178
Chapter 5: Risk Reduction Strategies
5.1. Overview 181
5.2. Policy Methods and Approaches 183
5.2.1. Regulating Social Media Platforms 183
5.2.2. Improving Disinformation Education and Awareness 186
5.2.3. Establishing Norms for Machine Learning and Autonomous Systems 188
5.3. Technical Tools 190
5.3.1. Developing AI-enabled Detection Tools 190
5.3.2. Leveraging Authentication Tools 192
5.3.3. Establishing Transparent Communication Channels 193
5.3.4. Improving Explainability of AI-enabled Tools 194
5.4. Arms Control and Risk Reduction 196
5.4.1. Declaratory Policy 196
5.4.2. Eliminating Dual-Capable Delivery Systems 201
5.4.3. Mitigating the Risks of Entanglement 202
5.4.4. Eliminating Low-Yield Nuclear Weapons 204
5.4.5. Providing Reassurances on Second-Strike Forces 205
5.4.6. Negotiating Limitations on Hypersonic Weapons 205
Chapter 6: Scenario Exercise
6.1. Overview 209
6.1.1. Move One 214
6.1.2. Move Two 218
6.1.3. Move Three 222
6.1.4. Scenario Reflection 227
Appendix: Additional Resources
Risks of Emerging and Disruptive Technologies 228
Cuban Missile Crisis 230
Nuclear Deterrence and the Risks of Nuclear Weapons 230
Artificial Intelligence, Machine Learning, Autonomous Systems, and Deep Fakes 232
Cyber Weapons 234
Electronic Warfare 235
Hypersonic Weapons 235
Sensors and Intelligence, Surveillance, and Reconnaissance (ISR) Platforms 236
Social Media and Disinformation 237
Quantum Computing 238
Direct inquiries to: Andrew Facini, afacini [at] csrisks.org