With the quick advance of autonomous vehicle (AV) technology, understanding the potential demand of AV and its user characteristics has increasingly become a popular area of research. In consumer choice and technology adoption literature, whenever the demand of a new product is forecasted, the attitudes towards existing choices are found important in addition to new product attributes and consumer characteristics. While there is an abundance of literature from stated preference (SP) surveys identifying attitudes are just as important as demographics in forming a purchase or usage decision of AVs, past studies have seldom looked at how attitudes towards existing travel modes affect the new mode adoption. We conduct a dynamic online SP survey in Singapore on 2,003 individuals, with indicator questions about impressions on existing modes. We focus on how these attitudes affect AV adoption based on confirmatory factor analysis and discrete choice models with latent variables. The results show that having positive attitudes towards public transit casts a negative effect on AV adoption, while having positive attitudes towards ridesharing is positive on AV adoption. And, positive attitudes towards walking and driving do not have any significant effects. In addition, the model identifies that highly educated, wealthy, and/or younger people as the population to have more positive attitudes towards new technologies and more likely to adopt AVs. The research provides insights on potential relationship between AVs and existing modes, as well as the characteristics of potential audience, which may be of help in planning future AV services.