Toward Usable Theories for Human-Automation Systems
Résumé
A number of the engineering challenges raised by interactive software in the last decades have been addressed or worked around. When planning the development of large systems, user interfaces are often considered a solved problem and resources go elsewhere. However, it can be argued that some of the currently identified challenges (role of humans in large systems, explainability of AI) are reformulations of problems that are well known to the community of interactive systems engineering. We attempt here to distinguish which problems have been addressed by empirical methods and which remain to be solved by scientifical methods. We then propose requirements for theoretical models that will allow engineers to go beyond empiricism and increase their control over the systems they design and develop: engineers need models that consistently describe socio-cyber-physical systems, that are usable to address practical design questions, and that allow to automate part of the design and verification processes.
Origine : Fichiers produits par l'(les) auteur(s)