In this FrAIday talk Marinus and van den Houten share how hacking and tinkering can reveal cracks in the polished surfaces of generative AI. These cracks expose hidden power structures and extraction processes. Putting a mirror to the algorithm is a method of using feedback loops to reveal parts of deeper faults and biases in the algorithms. Exposing these parts of the inner problematics that would otherwise be smoothed over through slick commodification.
They will dive into their recent works to show how immersion and alternative storytelling open up new ways of engaging with AI’s complexities. One of these works, Model Collapse, depicts the landscapes from which AI emerges—its self-consuming origins in the extraction of earth’s resources and data—through a playful, glitching, ever-changing generative environment and archive of stories.
Welcome to join us for this #FrAIdaytalk at UmArts Research Studio or online!
To find out more about the FrAIday talks at Umeå University visit: https://www.umu.se/en/research/our-research/features-and-news/artificial-intelligence/fraiday/