kennedy+swan
Germany
Artist Statement
Description
What happens when sensitive health data is used to train medical AI models? None of the current systems are error-free. They hallucinate and tend to provide plausible diagnoses instead of admitting insecurities. One such hallucination is at the center of kennedy+swan’s work The Neverending Cure. An AI model specialized in recognizing lung cancer in tissue scans is submitted to a both playful and revealing experiment. Assisted by doctoral researchers at the BIFOLD AI Institute, kennedy+swan expose the algorithm to watercolors on glass – painted in the aesthetics of microscopic specimen. What looks like fine structures of real lung tissue to the human eye is recognized by the AI application as potentially diseased tissue and diagnosed as such with astonishing confidence. In their artistic research, kennedy+swan reflect on the blind spots of medical AI systems. How far are we from a diagnostic AI that we can trust wholeheartedly? From the AI analysis of the images to the complete immersion in the watercolors’ “cell structures,” we get closer to the feeling of opening up the body to artificial intelligence. Today’s challenges could provide crucial impulses to deliver eternal healing – fed by ever-new data that will liberate us, step by step, from our physical suffering. The Neverending Cure.
Process
As kennedy+swan, we received the scholarship of Berlin AI Institut BIFOLD and the Schering Stiftung (a foundation empowering the crossover of arts and sciences). The PhD students and scientists of BIFOLD helped us gain deep insights into cutting-edge healthcare AI models and granted us access to play with and scrutinize these tools, which might one day decide over life and death. Yet, delving into this tech-knowledge inspired us to create truly analog paintings on glass, starting a lively dialogue between artists and algorithms.
Tools
We ran the medical AI models on Google Colab notebooks. The code is written in Python. The cancer models are called "Virchow" and "Prism". The researchers helped to modify the codebase so that it accepted the scans of our paintings as input data. We used a laser cutter to etch the diagnoses, diagrams, and heatmaps of the AI into our watercolor paintings. Self-made lightboxes show the microcosm of the paintings in incredible detail.
