Images have always been complex and unruly elements: difficult to examine, to critique, and very open to different meanings. Still, images have often been treated as persuasive, honest representations of reality. AI now holds a magnifying glass to this issue, as convincing pictures can be created without cameras and any connection to the physical world. Generative AI is thus an opportunity to discuss larger questions: What happens to authenticity when we cannot “trust our eyes”? How should we look at images? How can we boost visual literacy, which has become a core skill for navigating current media environments? This talk focuses on how Generative AI highlights an essentially old problem—but in a new, urgent way.