Defining and Designing Reliable AI-Enabled and Autonomous Systems

The desire to have a reliable system is common across stakeholders. What is less common however is a true understanding of how reliability should be defined – especially when considering AI-enabled and autonomous systems. This paper intends to discuss the key elements of the definition of reliability and how that may impact the design of AI-enabled and autonomous systems. Understanding the key elements of how reliability is defined and measured will lead engineers to design and develop systems to mitigate potential failure modes and successfully deploy capabilities to accomplish mission objectives.