Memoria Fluminis is a narrative video work about a river, a silent witness to the history of humanity. Like the riverbed that collects the remains of past civilizations, the work carries fragments of our Collective Unconscious - as it emerges from the monumental quantities of data that AI has learned to process and imitate.
Like a memento mori on the scale of a river that will engulf our dreams as well as our anxieties, Memoria Fluminis invites viewers to step aside, to contemplate their place in the grand narrative of existence; far removed from the frenetic pace of our lives, the continuous stream of novelties and daily news, this contemplative work embraces broader time scales that go beyond us.
The story unfolds in the (Latin) words of the river itself, a timeless testament to the grandeur and folly of human endeavor, evoking both the building of gigantic empires and the intimate moments of quiet lives. It also underlines our profound interconnection with the environment: the scars of our wars, industries and pollution remain etched on the landscape, testifying to the lasting consequences of our actions. To leave us with this persistent question: will we learn from the whispers of the past, or repeat its echoes in an uncertain future?
The work takes advantage of the power of AI, a technology that is itself constantly evolving as new iterations emerge. In the struggle between two algorithms – one that imagines, the other that drips pixels – the eternal confrontation between entropy and negentropy is replayed. Between dreams of grandeur and the forces of oblivion, in the eternal cycle of excess and its consequences, Memoria Fluminis encourages us to listen to the whispers of the past and embrace the flow of existence.
Memoria Fluminis is part of the Clinamen series, a project that explores the depths of the Artificial Unconscious in flows of pixels, offering a journey through the labyrinthine corridors of the Artificial Unconscious.
The images were produced using custom AI algorithms, based on DeepDream and CLIP by OpenAI, as well as OpenGL shaders and procedural post-processing techniques.
The text was written and translated to Latin with suggestions from Google Gemini, and from open-source and custom AI models. The voiceover was generated with Microsoft Azure TTS.