Abstract Embodied robotic neurotechnologies such as medical exoskeletons or neuroprostheses can enhance the ability of neurological patients to lead independent lives and thereby increase their functional agency. Modern approaches leverage methods from artificial intelligence to detect neurologically generated intent and to automate complex parts of the actions based on contextual data. However, the potential broadening utility and apparent benefits of these neurotechnologies to restore, sustain, or expand individual agency can also incur risks of their overuse. Excessive automatization of movements could reduce the causal influence that individuals exert upon their actions and thus may compromise their sense of agency. The construct of a sense of self and thus an identity to which responsibility can be attributed requires that a person is able to distinguish self-generated actions from actions of others, attribute their actions to themselves, and take ownership of the actions’ consequences. However, when parts of one’s movements are assisted or automated by intelligent robotic neurotechnologies, the distinction between endogenous and exogenous actions becomes blurred, resulting in a co-agency. In this paper, we discuss how this co-agency between humans and neurotechnologies can be shaped through the lens of the cross-cultural value of harmony, emphasizing the central role of intentions, balance, synchrony, and context. We argue from a cosmopolitan neuroethical perspective that embodied robotic neurotechnologies should prioritize supporting their users’ sense of agency, next to and above increasing their capacity for objective functional actions. Optimality should not be the prime goal of assistive neurotechnology, but a harmonious balance between human and machine agency.