Dot-Tod and ModMatrix
Exploring the idea of decay:
With this project, we wanted to use artificial intelligence as a space for experimentation, play, and exploration of its boundaries. To do this, we pursued two completely different concepts.
With the AI model DOTTOD, we aimed to investigate how artificial intelligence perceives the concepts of degradation and trash through image generation. We provided input words to explore how the AI understood the concept and represented it in the various images it generated.
We used a completely blank image to observe the results without any visual reference to base them on.
Another avenue of exploration was experimenting with ModMatrix.
Concept of the hood => "En el hood no hay amigos, solo hay trash" (In the hood, there are no friends, there's only trash) - Yung Beef
We once again experimented with generating images related to phrases and quotes from trap songs, mostly in Spanish. We utilized an existing image and observed the various interpretations generated based on the quote. Many of these quotes are tied to street culture, leading to some containing "inappropriate" content, which prevented their inclusion due to site policies.
Reflection:
For future projects, I would like to continue investigating the biases inherent in artificial intelligence, starting from the prompts to study semiotics.
During that week, I, along with Ever and Nicolo , attended a talk given by @nonsenseBani at the CCCB on artificial intelligence. He stated that to speak a language, it was not necessary to understand it; it was enough to put together words that sounded good together. To illustrate this point, he mentioned that in one of his classes, he asked his students to verify the references of a bibliography he had generated using chat GPT. The bibliography was very well structured, listing the names of articles, recognized artists, and thinkers, etc. Of the 10 references provided, he said, 9 were invented. Some of the titles did not exist, others had mixed the author with an article from another author, and at other times, the titles did not match, etc.
Concepts I would like to delve into for future research include:
- Copy
- Original
- Reappropriation
- Property
- Authenticity
- Physicality
- Trust
- Prompt = poetry ?
- language
Furthermore, I have previously worked with image generation AI. Having studied digital arts, I have implemented such tools almost since the beginning of my career. I enjoy integrating them into my creative process not only to facilitate and automate tasks but also for visual research.
New AIs I would like to implement and play with:
- Text to 3D + 3D printing
- AI + Render in Blender
Link to see our presentation https://drive.google.com/drive/u/4/folders/1J7Ke-FrWW5kEcb-BRCbV9X2uFY0xXNso
Dot-Tod and ModMatrix
Exploring the idea of decay:
With this project, we wanted to use artificial intelligence as a space for experimentation, play, and exploration of its boundaries. To do this, we pursued two completely different concepts.
With the AI model DOTTOD, we aimed to investigate how artificial intelligence perceives the concepts of degradation and trash through image generation. We provided input words to explore how the AI understood the concept and represented it in the various images it generated.
We used a completely blank image to observe the results without any visual reference to base them on.
Another avenue of exploration was experimenting with ModMatrix.
Concept of the hood => "En el hood no hay amigos, solo hay trash" (In the hood, there are no friends, there's only trash) - Yung Beef
We once again experimented with generating images related to phrases and quotes from trap songs, mostly in Spanish. We utilized an existing image and observed the various interpretations generated based on the quote. Many of these quotes are tied to street culture, leading to some containing "inappropriate" content, which prevented their inclusion due to site policies.
Reflection:
For future projects, I would like to continue investigating the biases inherent in artificial intelligence, starting from the prompts to study semiotics.
During that week, I, along with Ever and Nicolo , attended a talk given by @nonsenseBani at the CCCB on artificial intelligence. He stated that to speak a language, it was not necessary to understand it; it was enough to put together words that sounded good together. To illustrate this point, he mentioned that in one of his classes, he asked his students to verify the references of a bibliography he had generated using chat GPT. The bibliography was very well structured, listing the names of articles, recognized artists, and thinkers, etc. Of the 10 references provided, he said, 9 were invented. Some of the titles did not exist, others had mixed the author with an article from another author, and at other times, the titles did not match, etc.
Concepts I would like to delve into for future research include:
Furthermore, I have previously worked with image generation AI. Having studied digital arts, I have implemented such tools almost since the beginning of my career. I enjoy integrating them into my creative process not only to facilitate and automate tasks but also for visual research.
New AIs I would like to implement and play with:
Link to see our presentation https://drive.google.com/drive/u/4/folders/1J7Ke-FrWW5kEcb-BRCbV9X2uFY0xXNso