In the gig economy, workers complete one-off tasks for compensation via digitally-mediated, on-demand labor platforms. While demographic data indicates that many gig workers are economically precarious and belong to marginalized social groups, little is known about how marginalization impacts their experiences. To fill this gap in research, this dissertation presents two qualitative studies on the opportunities and risks presented by the gig economy for workers who are economically and socially marginalized. The first study examines how workers navigate privacy-related risks in crowdwork in light of their economic dependence on such work. This study demonstrates that economic considerations are a key factor in how workers decide whether or not to complete privacy-invasive tasks, and problematically, that financial need can compel workers to complete risky tasks despite their privacy concerns. Further, I show that managing privacy in crowdwork constitutes a significant amount of unpaid, invisible labor that must also be shouldered by workers. The second study focuses on workers with disabilities and their experience with a wide variety of gig work, including ridesharing, delivery services, freelancing, and crowdwork. This study finds that gig work can offer disabled workers several benefits that are lacking in traditional employment, but that they also face several disability-related challenges in accessing and completing work. Navigating these challenges involves a large amount of invisible labor, and workers' dependence on income from gig work can further disenfranchise them and restrict their ability to exercise discretion in the tasks they choose to complete. Together, this dissertation identifies a key tension: gig work can be a vital source of needed income for workers who face broader economic and social exclusion, but at the same time, marginalization orders and shapes several aspects of how workers complete and experience gig work. I show how the power asymmetries on labor platforms and broader marginalizing forces combine to present workers with both new and amplified risks and challenges. Further, workers have to perform a significant amount of invisible labor to mitigate these challenges, and I show how the burden of this labor is acutely felt by workers who are already marginalized. I also call attention to how many workers can face complicated intersectional challenges based on multiple marginalized identities, such as along the lines of race, gender, sexual orientation, disability, and socioeconomic status. This dissertation makes three main contributions. First, I provide an empirical understanding of the opportunities and challenges that come with working on digitally-mediated labor platforms for workers who are economically precarious, have one or more disabilities, and/or hold multiple marginalized identities. Second, I add to theoretical understandings of the power asymmetries on labor platforms and their consequences, focusing on the impact on marginalized workers. Finally, based on my findings, I put forth several suggestions for how labor platforms can be designed to be more inclusive for workers with diverse needs and backgrounds.