<p>Carebots are robots developed to assist users within healthcare contexts. The current literature on carebots and their impacts on autonomy has standardly discussed only some components of autonomy rather than offering a comprehensive analysis. In this paper, we propose a more thorough framework. This framework applies Catriona Mackenzie’s multidimensional account of autonomy to factors in human–robot interactions that significantly affect personal autonomy. We show, through examining existing carebots with this framework, that their impacts on autonomy are indeed multifaceted. We also argue that their impacts are dynamical in structure, where promoting some conditions of autonomy results in diminishing others or even at the expense of others. One implication of our analysis is that this framework can and should be applied to assess the impacts of carebots upon autonomy at the designing phase, as well as when we consider adopting their use in different contexts.</p>