Strategic competition in the age of AI: Emerging risks and opportunities from military use of artificial intelligence

Table of Contents

• Background to this study 3
• About RAND 3
• Preface 3
• The advent of AI is ushering in profound changes to competition and conflict 4
• Summary 4
• AI poses complex significant and underappreciated risks to defence and security 5
• Equally leadership on military AI could deliver outsized benefits 6
• Urgent action is needed to mitigate emerging risks and exploit opportunities 6
• Table of contents 8
• Tables 10
• Figures 11
• Boxes 12
• Abbreviations 13
• Acknowledgements 15
• Chapter 1. 17
• Introduction 17
• 1.1. Research scope objectives and methodology 17
• Chapter 2. 19
• Towards a framework for strategic impacts from military AI 19
• 2.1. Problems with current understanding of the impacts of military AI 20
• 2.2. Contours of a possible conceptual framework 23
• 2.3. Summary 27
• Chapter 3. 28
• Impact at the national level 28
• 3.1. Understanding strategic advantage 29
• 3.2. AI impact potential for advantage 30
• 3.3. AI impact propensity for advantage 37
• 3.4. Summary 46
• Chapter 4. 49
• Impact at the international level 49
• 4.1. Understanding the international system 50
• 4.2. AI impact actors goals and power 50
• 4.3. AI impact global governance 52
• 4.4. AI impact strategic competition dynamics 53
• 4.5. Summary 58
• Chapter 5. 59
• Implications by competition type 59
• 5.1. Implications for alliances and partnerships 60
• 5.2. Implications for Defence engagement and capacity building 61
• 5.3. Implications for resilience and emergency preparedness 62
• 5.4. Implications for sub- threshold operations 63
• 5.5. Implications for deterrence 64
• 5.6. Implications for crisis management 68
• 5.7. Implications for conventional warfighting 68
• 5.8. Implications for nuclear warfighting 70
• 5.9. Implications for de- escalation peacebuilding and reconstruction 71
• 5.10 Summary 72
• Chapter 6. 73
• Implications by actor type 73
• 6.1. Implications for different types of state 74
• 6.2. Implications for different systems of government 82
• 6.3. Implications for non-state actors 84
• 6.4. Summary 87
• 7.1. Towards a prioritisation of strategic risks and opportunities 89
• Chapter 7. 89
• Priority issues 89
• Chapter 8. 93
• Lessons from other domains 93
• 8.1. Existing models of risk management 94
• 8.2. Transferrable learning 98
• Chapter 9. 105
• Toolkit of measures to exert influence 105
• 9.1. Mapping risks and opportunities against the toolkit 106
• 9.2. Mechanisms to boost AI adoption and benefits for Defence 106
• 9.3. Mechanisms to restrict AI adoption and benefits for non- state and terrorist actors and hostile and rogue states 111
• 9.4. Mechanisms to shape and influence governance arrangements 114
• Given gaps in understanding there is an urgent need for awareness raising problem finding and information sharing about military AI risks 115
• Governments should develop transparency and confidence-building measures with both allies and competitors to build trust and reduce escalation risks 116
• Governments should promote an inclusive participatory approach to build global norms of responsible behaviour around military AI as a prelude to formal agreements 117
• Governments should in parallel develop minilateral mechanisms for reducing nuclear- and bio-related AI risks as an urgent priority that cannot await global consensus 118
• Governments should investigate ways to incorporate AI tools into verification and compliance mechanisms and vice versa 119
• States should over time aim to consolidate the current fragmented landscape of AI governance initiatives into a more concrete architecture 119
• 9.5. Summary 120
• Chapter 10. 121
• Conclusion and next steps 121
• References 123
• A.1. Research approach 139
• A.2. Data collection methods 139
• Annex A. 139
• Methodology 139
• Annex B. 142
• List of interviews 142