Ex|po|si|ti|on. English: ex|po|sure. From latin "ponere"; to set up, erect, display, expose (to dangers), expose to public, abandon, terminate. Meanings: 1. public display / 2. introductory part of a story or drama / 3. exposure of an organism to environmental influences. To be unprotected/exposed to: attention/possible attacks/dangers due to spatial location, personal situation, or similar / 4. exposition (photography) / 5. concluding part of a confrontation therapy (encounter with fear-inducing stimuli).
In his solo exhibition Exposition, Christian Sievers presents an interactive system in which the audience can interact with two artificial intelligences and expose themselves to their decision-making and predictive capabilities. In a "recording room", an AI-based visual scene recognition system analyzes and describes the events in front of a camera. The audience entering this recording room sees themselves in the video image along with a description of the scene written by the AI, as direct feedback on their own actions in front of the camera. The scene recognition not only registers individuals or objects, but makes assumptions about how people interact in front of the camera. In case of certain common actions of the audience a second process is triggered. Another artificial intelligence, specialized on completing stories, is being given the information about the situation in the art space and is asked to imagine what could go catastrophically wrong. The audience becomes an actor in automatically generated narratives in which a nascent technology formulates our fears and desires towards itself.
Especially the big AI models like those used here, Azure and GPT-3, are expressions of a kind of collective psyche. The installation provides insights into how these models work. At the same time, an algorithmically monitored space provokes the audience to exuberance, something that does not come easily considering the well-known manifold problems around data protection, "bias" and the abuse of power by Silicon Valley. In the exhibition space, the installation replicates the digital decision-making architectures familiar from platform capitalism: To participate, one must suspend and relinquish control; a cognitive dissonance we all experience daily online.
The generated dark and hilarious narratives are shown on the screens in the exhibition space and - depending on consent - on https://comfort.technology. All dynamically generated content is highlighted in green. In order not to let the algorithms have the last word, (human) assistants accompany the production and reception of the narratives. They can intervene in the narratives, edit them, weigh them, comment on them or delete them independently and in consultation with the people recorded. From their competence as technically skilled artists, they mediate between the artificial intelligences and the audience, explain the technical background and provide access to the discourse on the power of algorithms.
An art project by Christian Sievers, implemented by Valia Fetisov. Graphic Design by Florian Egermann. Video documentation by Damian Weber. Assistance and data curation by Ting-Chun Liu, Kim, Sakaya, Jasmina, Juan, Giorgi, Sina and others. Visual Scene Recognition by Microsoft Azure Cognitive Services, Region Germany West Central. Natural language processing by GPT-3.
