We present a series of artworks created using Autolume, a no-coding, generative AI tool that leverages the artistic potential of Generative Adversarial Networks (GANs). Designed to empower artists without technical backgrounds, Autolume enables small-data model training, real-time latent space exploration, and interactive applications such as audio-reactive visuals. By preserving the fluidity and navigability of GAN latent spaces, Autolume offers unique possibilities for the creation of moving images and dynamic visual experiences. Through diverse case studies, including audiovisual performances, interactive installations, and small-dataset model crafting, we demonstrate how Autolume expands creative agency and supports new modes of artistic expression with generative AI.