Artificial intelligences already outperform humans in some fields such as playing Go or identifying tumors within X-rays. They often excel by developing whole new ways of seeing, or even thinking and recent findings suggest that they attain superhuman performances even without the need of learning from human experiences. AI is set to play an increasingly significant role in our professional and personal lives with applications within financial businesses, communication, medicine or autonomous cars. But as machine learning becomes more and more powerful and complex, the field’s researchers increasingly find themselves unable to account for what their algorithms really know. We experience a transition from causality, the basic foundation of the enlightenment, to an opaque world of correlations and statistical probabilities. The impact on our society is huge and this raises a number of questions: Just how much can and should we trust these intelligent systems? How much responsibility are we prepared to hand over to AI? How should we respond to the steadily growing capabilities of many areas of deep learning? Society as a whole needs to have a broad-based discussion on the applications and implications of AI.