GMU:Hallucinating computers and dreaming non-machines/Lina Louise Wolff

Further down the rabbit hole// Project by Alaina Nugnis& Lina Wolff

This work is based on a previous semester project witch explores how the lack of female representation in research and clinical trials has led to disproportionate adverse drug reactions in women. These effects are the result of healthcare systems historically designed around male bodies, often treating them as the default in diagnostics, treatment, and pharmaceutical development, showing that care systems are taking better care of male bodies then of any others.The unequal treatment gets visible through gaps, most common known as the gender data gap.

The installation focuses on visualizing the blank which is created through data blind spots and takes a closer look on how AI fills these blanks. AI s function by analyzing visual input through multiple neural network layers. With each layer, specific features are exaggerated, leading to a progressively surreal image. When the algorithm encounters ambiguity, it begins to “hallucinate”, projecting its own learned patterns onto the input. This work critically examines how both care systems and AI technologies are structured around existing assumptions, ultimately asking: whose bodies are being seen, and whose are being overlooked? In an analoge way the artwork reinterprets how image dreaming AIs, like Google’s DeepDream, start to hallucinate and dream and therefore tries to close gaps.

The artwork is a combination of analoge as well as digital parts. Using plant structures that are explored in the BioLab and analyzed under the microscope the gabs get visible through an abstract approach. Multiple layers of glass hanging right behind each other represent the structure of AI systems.

Background information

This work is based on a previous semester project witch explores how the lack of female representation in research and clinical trials has led to disproportionate adverse drug reactions in women. These effects are the result of healthcare systems historically designed around male bodies, often treating them as the default in diagnostics, treatment, and pharmaceutical development, showing that care systems are taking better care of male bodies then of any others.The unequal treatment gets visible through gaps, most common known as the gender data gap.

Idea/Concept

The installation focuses on visualizing the blank which is created through  data blind spots and takes a closer look on how AI fills these blanks. AI s function by analyzing visual input through multiple neural network layers. With each layer, specific features are exaggerated, leading to a progressively surreal image. When the algorithm encounters ambiguity, it begins to “hallucinate”, projecting its own learned patterns onto the input. This work critically examines how both care systems and AI technologies are structured around existing assumptions, ultimately asking: whose bodies are being seen, and whose are being overlooked? In an analoge way the artwork reinterprets how image dreaming AIs, like Google’s DeepDream, start to hallucinate and dream and therefore tries to close gaps.

Tech implementation

The artwork is a combination of analoge as well as digital parts. Using plant structures that are explored in the BioLab and analyzed under the microscope the gabs get visible through an abstract approach. Multiple layers of glass hanging right behind each other represent the structure of AI systems.

References

Artists:

Ryoji Ikeda-> sound artist

Kristen Jakubek