If you want to start with AI, please remember: Your AI is as good as your data is! Ofcourse you knew, but this presentation will open your eyes, even when you thought they were already open.
AI systems are expected to increasingly take autonomous decisions, often well-hidden from view. This development is not necessarily without danger. In this presentation Jaochim Ganseman will give you a high-level overview of some common pitfalls in AI-powered applications: snares and snags that require more thought than many AI gurus would admit. On the menu: bias and fairness, confounding variables, adversarial attacks, ethics, explainability and interpretability, and more!