A growing number of cyber-physical systems (CPSs) application domains, such as transportation, energy, sustainability, health, and disaster response, involve humans in nontrivial ways. An emerging application paradigm along with this trend is the use of humans as sensors, which is also commonly known as social sensing. A critical challenge in social sensing is that data sources may be unreliable: a common thread in CPS research focusing on the reliability of CPSs. Current research has mostly focused on the correctness of temporal behavior and correctness of software function. In order for social sensing to become a viable component in CPS feedback loops, it is crucial to understand the correctness of collected observations from unreliable individuals as well. We call this latter challenge the data reliability challenge. This chapter will review recent progress and the state-of-the-art techniques developed in various communities to address the emerging data reliability challenge in CPS.