Journal of Civil Engineering and Architecture 15 (2021) 229-246 doi: 10.17265/1934-7359/2021.05.001 Project Risk in the Age of Artificial Intelligence: A Metanetwork Assessment of Behavior-Centric Intangible Risks Christopher Cox 1 and Hamid Parsaei 2 1. Kimmel School of Construction Management, Western Carolina University (August 2021), Cullowhee, NC 28723, USA 2. Wm Michael Barnes ’64 Department of Industrial and Systems Engineering, Texas A&M University, 3131 TAMU College Station, TX 77843-3367, USA Abstract: Industrial projects can be viewed as complex sociotechnical systems (e.g., human agents interacting with technology) where cause-and-effect relationships do not necessarily occur in time-and-space proximity. For this work, metanetwork (e.g., a network of networks) analysis was applied to emergent behavior-centric intangible risks (BCIRs) in a portfolio of projects in the energy sector. A user-friendly framework is proposed to identify and quantitatively assess BCIRs, along with the conditions that initiate them throughout the project development cycle. The underlying hypothesis is a structured approach to identifying, assessing, and proactively addressing BCIRs that have the potential to improve a project team’s ability to meet its objectives. While we build upon Rasmussen’s dynamic safety model and address the need for a framework to assess causal factors that influence behaviors in the context of an energy-sector project, we do this with a view to a future where technology (e.g., artificial intelligence (AI), automation, robotics, etc.) will play an ever-increasing role. The proposed framework is presented as tested in a live project portfolio setting where organizational modifications were identified, simulated, and implemented. One particular dimension of the analysis, the issue of authority without responsibility, is also discussed. The results of this empirical assessment were further validated by an industry panel of subject-matter experts (SMEs). Key words: Intangible risk, project development, metanetwork analysis (MNA), sociotechnical system. 1. Introduction The emergence of automation and artificial intelligence (AI) technologies in the coming decades will transform the way people work and the skills required for effective performance. In addition to the need for technical skills, the need for social and emotional skills will increase substantially [1]. Developing a framework to address behavior-centric intangible risks (BCIRs) throughout the project development cycle can enhance risk management in this new era of human and machine interaction. In this paper, we focus on the BCIRs in an energy-sector project portfolio. The energy sector is the most Corresponding author: Christopher Cox, PhD, assistant professor, research fields: risk management, implications of AI on curriculum and retraining. capital-intensive and volatile industry in the world [2, 3], with 2017 capital expenditure reported at $714 billion [4]. The environment in which this investment takes place is one where stability and certainty are rare, with complexity and ambiguity dominating the landscape [5]. Projects can take many years to move from conceptual planning to initial operation, subjecting them to a myriad of risks. Empirically, 60% to 80% of projects fail to be completed on time and within budget [6-8], with these failures due in part to human behavior [9]. For instance, fewer than 25% of oil and gas projects undertaken in the North Sea between 2011 and 2016 were completed on time and within budget, and the Oil and Gas Authority (OGA) of the United Kingdom has attributed many of the reasons for projects not meeting expectations to events that are “non-technical in nature” [10]. D DAVID PUBLISHING