AI and autonomy are two closely linked concepts. While AI refers to computer-controlled systems, autonomy describes how independently they can act from human operators.
NATO has long understood the need to prioritize technological integration with AI and autonomy and published its first AI strategy in October 2021. Notable AI projects include the Chinese Brain Project’s mind-controlled drones, and Russia’s MiG system, which recommends maneuvers to fighter pilots.
Amongst Allied capabilities, the United States Project Maven utilizes target recognition systems to identify hostile combatants and platforms, and the British army’s Battlegroup Command and Control Trainer runs conflict simulations. For NATO to maintain its viability and relevance in the modern era, it is vital that the Alliance remains competitive in this arms race.
Our study was predicated on the need for a NATO-wide solution for AI and autonomous systems procurement, development, and scaling. But why should NATO be the venue for this change and innovation? Are there better options for such efforts? We had to first justify the focus on NATO as the instrument for innovation and change before we could research the problem and formulate our recommendations.
Three benefits of NATO came to mind.
- First, NATO sets the stage for promoting interoperability of both traditional and emerging technologies. As we discovered early on, AI and autonomous systems work best when built on networks or “swarm” structures, which are only possible with interoperability.
- Second, NATO allows allies to coordinate their own spending on research and development (R&D), which reduces wasteful duplication issues.
- Third, using NATO as the main forum for AI and autonomous systems ensures better and more holistic procurement and specialization within and between Allies.
Our team began the project with an initial problem-scoping assessment of the current threat environment and technology landscape. Recognizing that AI and autonomy are two distinct but interconnected concepts, we drafted an initial map while keeping in mind three core questions: What should NATO’s AI and autonomy strategy be? What policies will be critical? What capabilities are the most critical to address?
To investigate why NATO’s capability gap over adversaries is shrinking, we produced a mutually exclusive and exhaustive issue tree exploring current threats and possible future solutions. This aims to aid NATO maintaining its edge as AI and autonomy continue to grow more entrenched in each new emerging technology.
Throughout the study, the team maintained close and continued working relationships with both the NATO Innovation Unit in Brussels and the NATO ACT Innovation Hub in Norfolk, Virginia, to ensure that the project fulfilled the needs of major allied stakeholders and maintained relevant to NATO’s burgeoning strategy for AI and autonomy.
Following our scoping of the problem, we conducted an in-depth analysis of the current state of AI and autonomy in the defense sector, seeking an understanding of its uses and potential implications for future warfare. Although we were limited to publicly available information, we were able to gather a reasonably clear assessment of the current state of AI and autonomy within NATO and the potential threats it poses to the alliance.
To further understand where specific NATO capabilities stand in relation to peer competitors, we chose to investigate three case studies, each of a different class of AI, which compare a NATO capability to a non-NATO capability. Specifically, we explored wargaming systems, brain-computer interfaces, and target recognition capabilities while keeping autonomy as a common thread between cases. From these case studies, we concluded that at present, NATO appears to be leading the AI race but risks losing its competitive advantage to peer competitors if allies fail to leverage the private sector, coordinate implementation, and engage with the public.
Using these case studies and published NATO AI strategy goals, we divided our recommendations among five policy areas:
- Reasonable development
- AI literacy
- Private sector innovation
As we developed our detailed recommendations, we categorized each suggestion under one of the five policy “buckets”. In sum, we formed eight recommendations, assessing each on its ease of adaptability and level of impact. For a more granular perspective, our team chose three of these recommendations, formulating suggestions for one-, three-, and five-year performance indicators as a roadmap for implementation.
Ultimately, we hope that these recommendations enable NATO allies to better innovate, scale, deploy, and integrate AI and autonomy-based technologies to form agile, system-wide solutions. These new capabilities will revolutionize NATO’s military and strategic affairs, thus strengthening NATO’s ability to fulfill its essential core tasks of collective defense, crisis management, and cooperative security.
Nicholas Nelson – Center for European Policy Analysis, Georgia Tech Research Institute
Nico J. Luzum – Center for European Policy Analysis, Georgia Institute of Technology
Garritt Garcia, Anika Gouhl, Jack Sheldon, and Maria Winstead, Georgia Institute of Technology