A.M.I was a piece that unexpectedly shifted direction mid-creation. Built as a direct response to The Watsons, it continues an unwritten narrative documenting a significant period in my life. The piece explores the theme of self-representation and how we misrepresent ourselves through social media, often creating an idealized version of reality. These “faux-selves” hide imperfections and insecurities, but when examined closely, they reveal the subconscious ways we conceal parts of our lives—whether by ignoring certain aspects, living a pseudo-life, or suppressing emotions.
A.M.I embodies the overwhelming emotions that bleed into reality, represented as a physical manifestation of a brain grappling with memory processing. The piece explores how taking an overly rational, clinical approach to life can detach individuals from their emotions, creating a third-person perspective—akin to watching oneself in a reality TV show.
This was also a deeply personal piece, reflecting a challenging time in my life. Where The Watsons expressed the joy of finding a supportive group of friends and family, A.M.I delves into the darker side—paranoia and anxiety that often follow such connections.
I aimed to physically represent the overwhelming thought process, and the brain was a natural choice. Projection mapping was ideal for animating and visually guiding the user through the experience. I began experimenting with materials, initially trying lightweight expanding foam for the brains. However, the foam produced porous, unpredictable textures unsuitable for projection mapping.
I switched to a more durable approach: sculpting a clay model, creating a silicone mold, and casting the brains in plaster. This method allowed for precise hand-finishing and customization, ensuring each brain was unique. These organic differences represented individuality but posed challenges for projection mapping due to the lack of uniformity.
Above (left to right): A prototype foam brain, the silicone mold, and the final plaster casts.
The technical challenges of A.M.I required creative problem-solving. To simplify, I broke the digital system into three components: the front-end mapping system, the script builder, and interaction.
Projection mapping seven objects simultaneously, with dynamic content for each user, was particularly complex. To maintain a smooth frame rate above 30fps, I implemented a system that pre-downloaded the entire “payload” for each user over a reliable Ethernet connection.
I programmed the mapping system in C++ using the Cinder framework for OpenGL. By pre-rendering visuals and audio in JSON scripts, the system minimized real-time processing. This ensured a stable frame rate averaging 50fps on my 2015 MacBook Pro.
Above: The mapping mode within A.M.I’s front-end system.
The script builder, written in Python, PHP, and Shell, generated personalized JSON scripts for each user based on analytical data from their Twitter accounts. This data included URLs for images, pre-generated audio files, and timing information.
The system analyzed 300 recent tweets, using sentiment analysis and keyword extraction via RAKE. It also scanned 100 photos for faces with OpenCV, adding a personal touch to the experience by displaying familiar images during the projection.
Instead of relying on web-based interaction, I used text messaging via Twilio, which felt more personal and eliminated issues with poor internet reception. Users provided their Twitter handle, which triggered the script-building process. The system even included a queue notification, letting users explore the exhibition and return when their turn approached.
This interaction method created an intimate experience, drawing users into the piece and leaving them with a memorable conversation.
For A.M.I, intimacy was central. I aimed to overwhelm users with information through droning, monotone audio, rapidly changing visuals, and an anxious soundscape. These elements combined to immerse the user in a reflection of their own emotions, leaving them alert and introspective.
The response during the opening night was overwhelmingly positive. Many found the piece adaptive and deeply personal, particularly the integration of familiar images, which made them feel part of the artwork.
Creating A.M.I was both technically demanding and emotionally vulnerable, but its reception validated the effort and highlighted the importance of user connection.