This document reports on the Epinet workshop on the making of robot autonomy, held in Utrecht 16-17 February 2014. The workshop was part of a case study focused on developments in this area, in particular, autonomy for assistive robots in care and companionship roles. Our participants were of relevant expertise and professional experience: law and ethics, academic and industry robotics, vision assessment and science and technology studies (STS). The workshop was intended to explore the expectations of robot autonomy amongst our participants, against a backdrop of recent policy views and research trends that are openly pushing an agenda of "smarter", more dynamic and more autonomous systems (e.g. European Commission, 2008; EUROP, 2009; Robot Companions for Citizens, 2012). Robotics development is intimately connected with visions of robot autonomy, however, as a practical achievement, robot autonomy remains till this day part real, part promise. Ideas of robot autonomy are nevertheless powerful societally and culturally-specific visions, even if the very notion of "autonomy" is vague and inconsistent in recent accounts of future robots. These accounts still come together with considerable force in directing the efforts of researchers and experimenters, for example, in establishing funding priorities. They have a function in strategic planning for future developments. Accounts of future robots are also informing and shaping the efforts of legislators, ethicists and lawyers. To that effect, one can say that there is an official vision of future robots, a yardstick with which everyone implicated in robotics development has to measure their expectations.