State of the Practice Survey - Predicting the Influence of AI Adoption on System Software Architecture in Traditional Embedded Systems.

Artificial intelligence (AI) is a very disruptive technology. When adopted by a software system, AI influences and significantly changes its architecture due to its complexity, as well as due to a need to adjust the existing system to use AI (e.g., adopt accelerators). This is particularly critical in traditional embedded systems as they focus on a tight coupling of software and hardware. In this paper, we present results of a survey on how well companies in embedded software domain understand AI, how they perceive its possible benefits, and how they discuss the adoption of AI and its influence on their software system architecture. The goal of this survey is to evaluate architectural techniques that companies currently use when trying to assess the influence of adopting AI and to discuss the adequacy of these techniques for this task.