Many artifacts that software developers produce are written in natural language: code comments, commit messages, text in user interfaces, privacy policies, high-level description of the system, and so on. In this talk I will present my research work on using natural language processing (nlp) techniques to analyze such artifacts. We infer useful information such as pre- and post-conditions of procedures, and we use it to automatically generate test oracles. In this talk, I will give an overview of my recent contributions on using nlp to make software systems more reliable and trustworthy, together with my plans for the future.