In the last 5 years the field of generative modeling (also known as creative Artificial Intelligence), notably Generative Adversarial Networks (GANs), has progressed at an enormous pace. Generative modeling can be used to create convincing fake images and video footage (deep fakes), fake data that can fool and undermine the working of other AI systems (adversarial examples), make art, take over basic creative labor (production of video game graphics, interior design, fashion, music or recipes), produce synthetic (aka: simulated or imputed) data where real data are scarce/lacking, or anonymize data by transforming them into synthetic data. When is fake good and when is it bad? Creative AI holds promises and perils that raise fundamental questions deserving the attention of citizens, lawyers and policymakers. In this panel a lawyer, an AI artist, a computer science expert, a philosopher and a social scientist give their thoughts on the highly topical question of the role of creative AI in society.