AIMNESIA

Year  2019

Awards  Winner of the Stroom Encouragement Award

Exhibited  KABK Graduation Festival (July 2019, Den Haag)

NeverLand Cinema (July 2019, Rotterdam)

De School (August-September 2019, Amsterdam)

Photos  CĂ©line Hurka

Music  Pavel Dovgal - Music for meditation #1

Tags  AI, Media art, Graduation Project, Installation

The AIMNESIA project focuses on the concept of human hybrid memory, which can be augmented, influenced, and modified by AI. Through the experience of absurd video installation about a hybrid memory creation made by using a pre-trained BigGAN model, the Ganbreeder app, and a reverse Google image search, I intend to prompt a discussion of evolving algorithms that can be trained on our online photos and can fill in memory gaps by creating fake memories that are plausible enough to be perceived as real.

AIMNESIA consists of a speculative video installation about the machine processing of human memories and three-phase interviews with machine learning researcher and inventor of GANs Ian Goodfellow, futurist and tech-philosopher Gray Scott, professor of cognitive neuroscience Guillen Fernandez, computational neuroscientist Marco Aqil, artist and programmer Gene Kogan, artist and educator Tivon Rice, creative technologist Tomo Kihara, and creator of the generative design tools Joel Simon about the relationship between the human brain and AI, hybrid memories, and the future of GANs.

More at official website

Technological development is leading us to convert our minds into bits and to extend the memories, thoughts, and perceptions beyond our biological bodies to algorithmically mediated objects, databases, and networks. Our brains adapt to the tools that we use. The collapsing border between neuroscience and technology has led to the emergence of a new space wherein artificial intelligence (AI) and the human brain can learn from each other. Today, artificial neural networks play an unseen but crucial role in our digital ecosystem by defining and recommending what should be seen, listened to, and read. Within the current machine learning revolution, so-called generative adversarial networks (GANs) are trained on massive datasets of images, sounds, and texts and can generate realistic patterns.