This document represents a first step towards establishing and documenting Recognised Good Practice for the safety assurance of Autonomous Systems. It has been authored by the Safety of Autonomous Systems Working Group (SASWG). Consistent with the SASWG’s aims, the document focuses on novel challenges associated with autonomy technologies. This book represents the first release of objectives by the SASWG. It provides a small set of objectives, focused on the computation-level, that would be expected to be met in any compelling safety argument for Autonomous Systems. Whilst they are considered useful, it should be noted that these objectives have been developed from a theoretical basis; to date, they have not been subjected to practical use. This document focuses on Autonomous Systems that use Artificial Intelligence developed using Machine Learning. Although it is possible to envisage Autonomous Systems that do not use these technologies, they are considered to represent the greatest assurance challenges. As such, they have been the focus of the SASWG’s efforts.