I will argue that focusing on embodied agency is critical to explaining intelligence in biological and artificial systems, and therefore that robotics should have a more central place in AI research than it currently does. This includes even so-called “internal” processes such as perceptual belief-formation, reasoning, and decision-making. I will explain why neither the causal nor the normative properties of these processes (such as rationality) can be understood by analysis of computational systems alone, but must include substantive environmental factors. Lastly, I will show how a more embodied, embedded perspective on artificial intelligence helps us incorporate insights from feminist and social justice theorizing both in the design of AI systems and in understanding concerns related to AI fairness.