In the keynote, experiences and challenges around communicating the mathematics of core algorithms of AI, in particular Machine Learning, will be presented. How can the mathematics behind Neural Networks or Reinforcement Learning be explained? What is Gradient Descent, how does Q-learning work, and what are stochastic processes? We focus on open-source and interactive resources to enable a playful and self-discovery approach towards learning. School students and the general public can experiment with neural networks and train them themselves or guide a small learning robot through a 2D maze. We will show exhibits of an AI exhibition, an open curriculum for AI workshops and will go a bit deeper into a free MOOC on AI for educators with additional classroom material and mathematical explanations. At the end, we will also present a new project on AI and electronic music.