What is agency and why does it matter? In this work, we draw from the political science and philosophy literature and give two competing visions of what it means to be an agent. The first view, which we term mechanistic, is commonly—and implicitly—assumed in AI research. However, the mechanistic view is a fundamentally limited means to understand the ethical characteristics of AI. Under the second view, which we term active, AI can no longer be considered an agent. We ultimately argue that AI should be viewed not as an agent but as the outcome of political processes; this clarifies the questions of what ought to be built, and of how to seek accountability for harm.