Interactive Installation, DCGAN, Python, C++, 2023
This work challenges conventional boundaries between algorithmic creation and human sexuality. Using web scraping techniques and API manipulation, the project continuously harvests publicly shared intimate imagery from Reddit communities to train a Deep Convolutional Generative Adversarial Network (DCGAN).
Rather than simply reproducing explicit content, Meat Grinder exists in a deliberately ambiguous territory. The algorithm generates images that evoke sensuality without explicit vulgarity, questioning our cultural and technological definitions of what constitutes "pornographic" content.
The real-time training approach, where the dataset evolves with each new contribution to the public domain, positions the work as a living document of our collective digital sexual expression. Despite institutional resistance—necessitating multiple discussions with department administrators and university HR regarding the training data—this work stands as a bold statement on algorithmic bias, digital body autonomy, and the liberation of sexual imagery from traditional power structures.