While there is an increasingly large body of studies on the potential demand of autonomous vehicles (AV), one understudied factor is risk preference. Risk preference is important because many aspects of AVs are highly uncertain as the technology and its encompassing mobility system continue to evolve. This study analyzes how risk preferences influence the choice of AVs, and how the risk preferences elicited by the economic and the psychometric methods differ in their impacts. The authors conducted a stated preference survey in Singapore with 1,303 persons (7 choice scenarios per person). The authors extracted two economic risk preference parameters based on the prospect theory; and the authors extracted one psychometric risk preference parameter based on Likert scale questions using factor analysis. After applying mixed logit models, the authors found that risk-seeking preferences significantly increase the choice of AVs. The economic risk preference and the psychometric risk preference are orthogonal to each other and both contribute to predicting AV usage, and the economic risk measure improves the choice model more effectively. The two economic risk preference parameters are subtly different because one reveals information about the events with extreme-range probabilities, such as car accidents, and the other is more related to the events with middle-range probabilities, such as potential technical improvement of AVs. The results show that risk preference is an important factor influencing the future adoption of AVs, and future studies should continue to examine the relationship between the various uncertain aspects of the AVs and the various components of the risk preference.