Concerns about the societal impact of AI-based services and systems has\nencouraged governments and other organisations around the world to propose AI\npolicy frameworks to address fairness, accountability, transparency and related\ntopics. To achieve the objectives of these frameworks, the data and software\nengineers who build machine-learning systems require knowledge about a variety\nof relevant supporting tools and techniques. In this paper we provide an\noverview of technologies that support building trustworthy machine learning\nsystems, i.e., systems whose properties justify that people place trust in\nthem. We argue that four categories of system properties are instrumental in\nachieving the policy objectives, namely fairness, explainability, auditability\nand safety & security (FEAS). We discuss how these properties need to be\nconsidered across all stages of the machine learning life cycle, from data\ncollection through run-time model inference. As a consequence, we survey in\nthis paper the main technologies with respect to all four of the FEAS\nproperties, for data-centric as well as model-centric stages of the machine\nlearning system life cycle. We conclude with an identification of open research\nproblems, with a particular focus on the connection between trustworthy machine\nlearning technologies and their implications for individuals and society.\n