Autonomy is often considered a core value of (Western) society that is deeply entrenched in moral, legal, and political practices. The development and deployment of artificial intelligence (AI) systems to perform a wide variety of tasks—including driving cars, operating trains, assessing risk, or making recommendations—has raised new questions about how AI may affect human autonomy. It is unsurprising, then, that many guidelines on AI development emphasize the need to protect human autonomy. Yet, these guidelines also differ substantially both on what the risks to human autonomy are and on how we might address them. To systematically identify risks to autonomy, we first need to confront the complexity of the concept itself and then adapt governance solutions accordingly. In this talk, I argue that we need to consider distinct dimensions of autonomy---agency and authenticity---in order to systematically assess potential impacts of AI deployment on autonomy. Based on this distinction, I provide a first glimpse of the risk landscape and discuss several policy implications.