If a space telescope were able to hallucinate landscapes of a planet after collecting 12 TB of visual memories of Mars, what would those hallucinations look like? This real-time generative AI painting is a latent walk-through of the machine’s Mars data universe as it endlessly generates hallucinatory landscapes. The ever-expanding data universe of hallucinations not only represents the interpolation and synthesis of photographic Mars archives, but also becomes a cosmos in which dreams are the main currency of artistic creativity. Such machine-generated hallucinations allude to a connection with the vast cosmos and include visuals that trigger a sense of belonging to the Earth and to our immediate environments. In this work that never repeats itself, a real-time generative artwork, custom AI model trained by half million satellite images of Mars and processed real-time by the Studio’s signature fluid dynamics algorithm will be infinitely dreaming about Mars's surface. The artist’s exploration of AI data pigmentation and light through fluid solver algorithms accelerated by GPU computation and real-time ray-traced lighting manifests this inspiration by showcasing the most innovative methods available to AI-based media artists.

Machine Hallucinations: Space is an-going AI research of data aesthetics at the RAS Lab based on collective visual memories of space. Since 2016, Anadol and his team have been using custom algorithms to train machine intelligence in processing these vast datasets and unfolding unrecognized layers of our external realities. In this gallery, they invite the audience to reflect on the relationship between technology, AI, and space explorations by viewing how the machine intelligence hallucinates after “seeing” millions of space images taken by the Hubble, ISS, and MRO telescopes.

The artwork comes with an artist-signed 3D physical certificate with backup, a 75’’ digital canvas, a custom computer and software.