From self-driving cars and dating apps to ChatGPT, the increasing presence of AI has been widely championed – but there are limitations and risks, too. Is more data always a good thing? When do algorithms make better decisions then humans, when not? What is the difference between human and machine intelligence? How to tell realistic expectations about AI from marketing hype and techno-religious faith? Are we sleepwalking into surveillance? To answer these questions is helpful for staying in control in an increasingly automated world. In this talk, I introduce the stable-world principle, the adapt-to-AI principle, and the Russian tank fallacy. We need, now more than ever, to arm ourselves with knowledge about how to make better decisions in a digital age.