Eu vejo uma figura Gorda – Foto na Piscina
2025
In the nostalgic reflection of rescued images, time tells me there was so much there for me to see myself happy. But I didn’t see it. I saw a fat figure, out of place, wrong. And so I erased myself. Literally: I removed my body from the photo. Before doing that, I asked ChatGPT to describe the image as it saw it — without the filters of my pain. The description became a song lyric, and the music was generated by AI. Then, in the space where I once appeared, I used another AI to fill my absence with invented figures. The machine tried to remake me. Tried to guess who I was. Tried to fill the hole. But what it created were distortions, glitches, incomplete presences — and precisely because of that, they are touching. This work is an experiment with memory, body, self-image, and technology. A reflection on how we look at our past images, and how we now delegate to AI the impossible task of restoring what was once lived. If my real body was once rejected, what enters its place now? What kind of presence can occupy the space of an erasure?
AI Context
The artwork employs artificial intelligence on multiple levels: the original image description (ekphrasis) was written with the assistance of ChatGPT; this description was then transformed into song lyrics using the same tool and later set to music with AI Song Maker. The absence of the figure in the photograph was filled with AI-generated variations in MidJourney, creating artificial presences in place of the erased body. Here, AI acts as co-author, interpreter, and distorter of memory.
Karen Caetano
Brazil
Bio
Karen Caetano is a visual artist, photographer, and researcher. Her work explores memory, the body, image, and technology, focusing on processes of montage, artificial intelligence, and experimental visual practices. She holds a Master’s degree in Design, Art, and Technology from Universidade Anhembi Morumbi (São Paulo, Brazil) and develops projects that intertwine autobiography, affective archives, and a critical approach to visual normativity.