Jürgen Schmidhuber, pioneer of modern artificial intelligence and long-time advocate of self-improving learning systems, argues that intelligence, human or artificial, arises from systems that learn to predict and compress the world, and then use those models to create ever more intelligent behaviour. In his view, the universe may be seen as a computable system whose complexity and our ability to model it drive intelligence and discovery.

In this opening interview, Schmidhuber will explore how AI systems that model, simulate and compress reality fit into our evolving understanding of scientific method and truth in an age of simulations. Are these systems merely reproducing human biases via their models, or can they genuinely uncover deeper structure in the world? Join Jürgen Schmidhuber to discuss what it means for intelligence, creativity and science when our theories and simulations increasingly dominate our interaction with reality.