My doctoral research develops a deeper understanding of interpersonal trust online through computational methods, in the context of online exchange platforms including peer-to-peer marketplaces, sharing economy platforms, and social networks. Through analyzing images in product listings on eBay and LetGo.com, language in profiles on Airbnb, and networks in social groups on Facebook, I show how different algorithms help understand and predict interpersonal trust in each context. Findings reveal patterns of interpersonal trust. For example, high-quality images are perceived as more trustworthy than stock imagery; language of promises lead to higher perceived trustworthiness through conventional signaling; and smaller, denser, and more private social groups are trusted more. These findings inform the design of online exchange platforms. The algorithms predicting trust could also be used for better ranking and recommendation to “engineer” interpersonal trust. Going forward, I propose a lens of “networked trust” to view interpersonal trust online, which has three focuses: (1) cues in Computer-Mediated Communication; (2) embeddedness in social networks; and (3) increasing mediation by algorithms. The networked trust framework can be used to frame future trust research in other contexts, such as misinformation. Finally, two research agenda were charted by this dissertation — AI-Mediated Communication and AI-Mediated Exchange Theory, which future work can develop on.