Current technological change is rapid and far-reaching, more so than ever before in human history. It is transforming all dimensions of human life, leading to large-scale adaptation. Among the disruptive new technologies that are being introduced into society, social robots are distinguished by their hybrid existence between mere thing and mindful agent. They are physical machines capable of interacting with the surroundings and designed to collaborate with humans on human tasks while interacting in human ways. In contrast to rescue, delivery, and patrol robots that replace human labour, these robots take social roles such as tutor, peer, learner, companion, or assistant. On a more general note, social robots are expected to work closely with humans—as partner, colleague, family, and friend. Yet, that a robot can act as a social entity, does not entail the robot constitutes a social being in its own right [1]. The question is whether social robots are capable of participating in and contributing to human social institutions such as healthcare, education, and economy; and if this is so, the follow-up question concerns what this may entail for society in the longer perspective. The emphasis of the present talk lies on the first question, the supposed contribution of social robots to social institutions. Raising a few queries concerning the ability of the BDI-paradigm and affective robotics to provide an adequate reply to this question, at least in its present formulation, I will briefly outline an alternative that lays down a new path in HRI, based in the notions of embodied, embedded, dynamic, and distributed cognition. I claim that these notions are particularly well-suited for designing social institutional forms of HRI, because they permit modelling the relevant cognitive processes as unfolding in the physical space that humans and robots share. The environment provides the resources for HRI such as artefacts, routines, and embodied social norms that simultaneously constrain and enable the emergence of novel institutional practices involving human and machine.