Technology increasingly employs unreliable systems as a central means to interpret input. Common examples include voice and text input in mobile devices or lane departure detection in cars. This reliance exposes a fundamental problem – people do not generally understand the underlying systems, and seemingly small system or human errors can lead to potentially disastrous consequences. While technical improvements partially address this, recent research in my group pursues a complementary approach through a better understanding of human interaction with, and new user interface (UI) technologies for, unreliable systems. In this talk I first present work that analyzed human behaviours around occasionally failing systems in text entry and gesture recognition and the insights that we gathered from this work. Then, I present a new text entry method that reduces errors caused by auto-correction and prediction algorithms. Moreover, I present a system that uses a brain-computer interface to sense user reactions to incorrect auto-corrects, which we use to trigger better system responses by offering different corrections. Finally, I close with an outlook for future work in this area.