Frog Frame

Sharp Objects - EmProps

As a long-form generative AI homage to Artemisia Gentileschi, Britney Spears' knife dance, and all other objectified, abused and eventually self-empowered weapon-wielding women of history, "Sharp Objects" blends Gentileschi's Old Master sensibility with Empress Trash's stylization of saturated color palette and her traditional oil painting methods.

In her own words:

Following the concept of my first Emergent Properties series, "Dynastic," "Sharp Objects" at the core pulls from Gentileschi's work, which I studied in depth. The working title for this series was Artemisia's Revenge, thinking about how I express rage at personal and social injustices through art, much like Gentileschi did. As someone who is also a rape survivor, I remember when I learned of her work and story, I felt for the first time, "wow, if she did this hundreds of years ago, I can maybe be an artist too, in spite of everything."

Now, over 15 years later and into my second year of utilizing AI in my creative practice, I have observed consistently biases in all AI models to at least some degree of how women are depicted (if they are depicted at all). Without prompting to be otherwise, the standard is rather emotionless, very thin, and predominantly Caucasian. In an attempt to push against these standards in datasets pulled from our media and art, the women in this series are prompted with my bias to be emotive, diverse, and naturally imbued in dramatic and sometimes dancerly poses with butcher knives.

The collection variable is EMOTIONS, to probe how the base model I use - Stable Diffusion SDXL - applies the prompt to feminine expression. Sometimes it works; other times, the results are completely incongruent or interchangeable. There are methods with other AI tools to achieve an expression sought on whatever face, no matter race or gender, however, they are considerably more technically advanced, making them not easily accessible to the average user at this moment.

I'm of the thought that all technology is inherently neutral; it isn't until humans interact with it that it takes on any form of bias or ethical questions arise. AI models are constantly learning and developing as we interact with them, feeding them more data with images and language we choose or are allowed to use. Additionally, as a baseline, there are centuries of a variety of biases in art and media that the AI models are pulling from, bringing the question many ask to the forefront, "What are the implications of past and current biases on future representations of the human experience with AI?"


"Sharp Objects" is a Long-form generative AI-art collection. 'Long-form generative' means each work is generated at the time of mint with no curation of the raw AI output from the artist besides from the prompts and variables.

This collection was created by Empress Trash for her first solo exhibition in Paris, through Superchief.


Sharp Objects
Long-form generative AI-art
Edition of 500
Made with EmProps OpenStudio
Stable Diffusion + Upscaler
2560x4096px PNG
February 22, 2024