Abstract
This paper focuses on the design and effectiveness of adaptive automation in Safety-critical systems. Our perspective is to focus on the human agent in the system as part of the Meaningful Human Control (MAS) project, pivoting on how humans will be accounted for in system design to enhance human performance. The division of tasks, decision authority, and the extent of automation are among the challenges introduced by new systems. Adaptive automation could balance function allocation between humans and machines to improve performance. It is therefore important to know if and how adaptive automation can be effective. This review paper aims to (1) define adaptive automation, (2) highlight important factors in adaptive automation design and application in interactive systems, (3) show when adaptive automation can be effective, and (4) highlight the design implications and gaps identified. This is part of a broader systematic literature review on the successful design principles in automation in the past 10 years. Following the PRISMA model and applying exclusion criteria, 14 articles were selected and thematically analyzed. The results showed that adaptive automation could potentially improve performance depending on the specific context and design boundaries that are discussed in this paper.