During the year of 2022 I've started investigating the use of deep learning techniques such as Generative Adversarial Network (GANs), Texto-to-Image and Image-to-Image models with the intention to explore how these algorithms are able to reveal subtle and relevant aspects of how we represent and perceive ourselves. This experiment was an essential part of my post doctoral research at UFG.
- Technologies
- Generative AI
- Categories
- Artificial IntelligenceGenerative
These generative model's algorithms are mirrors, microscopes and telescopes aimed at us.
In this piece, I trained an AI model with images of myself and started to generate fake "imagined" versions of me and see which distortions would emerge when extrapolating the AI capabilities.
By pushing the AI’s capabilities, I aimed to explore and amplify the distortions and deviations that emerge when the model extrapolates beyond its typical parameters.
The result is a series of synthetic portraits that challenge our understanding of self-identity, questioning the boundaries between human and machine-generated perceptions. Can a machine show me parts of myself that are still undiscovered?
The final outputs offers insights into the evolving relationships between technology and human identity, serving as a testament to the potential of AI to not only replicate but also reinterpret, reimagine and recreate.
