Trust Management for Artificial Intelligence: A Standardization Perspective

Abstract

With the continuous increase in the development and use of artificial intelligence systems and applications, problems due to unexpected operations and errors of artificial intelligence systems have emerged. In particular, the importance of trust analysis and management technology for artificial intelligence systems is continuously growing so that users who desire to apply and use artificial intelligence systems can predict and safely use services. This study proposes trust management requirements for artificial intelligence and a trust management framework based on it. Furthermore, we present challenges for standardization so that trust management technology can be applied and spread to actual artificial intelligence systems. In this paper, we aim to stimulate related standardization activities to develop globally acceptable methodology in order to support trust management for artificial intelligence while emphasizing challenges to be addressed in the future from a standardization perspective.