An Approach for Enterprise Architects to Analyse Opportunities and Constraints for Applying Artificial Intelligence in Military Transformations Juha Mattila 1 and Simon Parkinson 2 1 Aalto University, Helsinki, Finland 2 RMIT University, Melbourne, Australia juhakaimattila24@gmail.com simpar92au@gmail.com DOI: 10.34190/ECIAIR.19.045 Abstract: Military Forces are seeking strategic advantages promised by the latest advances in Artificial Intelligence (AI) trained with data collected from networked military things. In some countries like USA, Russia and China, the quest to develop advantages with AI has become the next arms race. Most visibly, the race is on in autonomous systems, but there are further ambitions behind what has been published. Enterprise Architecture (EA) practitioners are being asked where and how to implement AI for the best military benefits. Firstly, the answers would be naturally based on the availability of data, the computer performance and stability of processes and environment. Secondly, the answers may be sought within an organisation’s cultural features, its ability to adopt new technology, create explicit knowledge, and flexibility when facing unexpected. The existing EA frameworks provide tools to answer the first questions but, unfortunately, do not include models to analyse the evolutionary or revolutionary paths for the development of military capabilities. EA practitioners are facing questions like: What is the readiness of military enterprise to benefit from features of AI? How fast does a military organisation adopt the opportunities offered by AI? Which AI features may be the best to start the transformation? This paper introduces an enhanced EA tool that improves the success of EA practitioner in assessing opportunities and challenges of military enterprises applying AI in their transformations. We approach the challenge from pragmatic view using design research methodology to define the problem, find a tool for a solution, demonstrate it in a case study and evaluate results. The proposed Tool helps the EA practitioner in modelling the full stack of an enterprise from culture down to technology. Secondly, it helps EA practitioner to recognise opportunities for using the features of AI. Thirdly, the Tool illuminates the interrelated evolutionary forces between the structural layers and reveal possible challenges in transformation. Consequently, the EA practitioner can provide advice to military decision makers in implementing features of AI, avoiding the typical pitfalls of practice. Keywords: Enterprise architecture, Artificial intelligence, Military transformation, Evolution of system of systems, Adaptive architecture, Evolutionary theory 1. Introduction Artificial Intelligence (AI) stands out as a new silver bullet to further transform digitally enabled military forces (US Congressional Research Service, 2019) (State Council, 2017) (Chui, et al., 2018) (Allen, 2019). The enthusiasm should be tempered since no technology alone has ever revolutionised warfighting. Whereas, history shows that difference is made when applying the technology into the ways of organising and fighting (Hollenbeck, David, & Jensen, 2018). AI is a quickly developing field of technology, revealing opportunities for the military at least in the following areas of application: autonomous systems, man-machine interface, analytics, and image recognition (Kai-Fu, 2018). Since the opportunities of AI may be broadly applied within a military force, the paper analyses the military transformation from an enterprise architecture (EA) viewpoint of an open, socio-technical system of systems (Mattila, 2017) focusing on three core process areas of military affairs: force utilisation, generation and support (Smith, 2008). Military organisations have a cultural heritage of being robust, specifically, resistant to disruptive outer forces (Levchuk, et al., 2001). Similarly, the intentional transformation of military enterprise will face resistance to change. Military forces that are in the focus of this paper are keen on exploiting the opportunities of AI features at least to keep up with the development of their adversaries or gain a strategic advantage but, excluding the autonomous weapons, the application of AI may be complicated because of the interdependencies of a socio- technical system of systems. Transformation of military capabilities using features of AI requires an understanding of the need for data (Berman, 2013) in training AI and the ability of an organisation to change its processes (Kegan & Lahey, 2009). Furthermore, there is a need for alignment between exploiting new technology, aspirations of military strategy and the actual implementation of transformation (Whittle & Myrick, 2005). This paper studies how an Enterprise Architecture (EA) practitioner can assess the readiness 215