m-onz


the end of art?

12.05.24




Does genai represent the end of art? At least as we know it? Or is it just another tool that opens up new avenues of creative exploration... I think its the latter.


I generated these images using a text prompt and text-to-image stable diffusion machine learning model running on replicate.com.


This is the exact prompt i used: "Imagine a vibrant, abstract art piece that captures the essence of Francis Bacon's iconic style, blending geometric shapes with bold, shocking colors. This artwork should evoke the vivid, intense emotions that Bacon's paintings are known for, drawing inspiration from his use of biomorphism and surrealism. Incorporate elements that reflect Bacon's fascination with the human figure, particularly his distorted, almost grotesque depictions that challenge the viewer's perception of reality. The piece should also hint at Bacon's interest in the interplay between painting and photography, perhaps by integrating photographic elements or techniques that mimic the movement and dynamism found in his work. Finally, the artwork should be set within the context of a modern art gallery, surrounded by the stark, clean lines of geometric architecture, creating a juxtaposition between the raw, emotional content of the art and the sterile, orderly environment of the gallery."


The negative prompt was "face, faces, legs, arms, humans, figures", I set it to 1024 x 768 and left all the other settings with their default values.


I used an LLM to generate that prompt using this prompt: "create and elaborate on this text-image prompt to more interesting artgeometric shapes bold interesting abstract colourful shocking vivid francis bacon painting in a gallery". Which takes this form of art creation laziness to new level. Prompt engineering machine learning models to make things seems is an interesting new method of generating media. I can also use an image-to-video ML to create extended videos from these images. I created an PoC of this here.


You can use a prompt like this with text-to-video or image-to-video ML models on replicate. If your trying to make video loops and content for a VJ set this would be a quick way to build up some material. I think that instead of "replacing" human creativity as some like to scare monger and sensationalise, AI is just another tool at an artists disposal. New creating processes and technology increases the power and potential output of artists who dare to see it as an opportunity in my opinion. Culture has stagnated for some time and I think that genai and automation can be a gateway to new and unexplored vistas of creative exploration for those dare to explore it.



stable diffusion : text-to-image
artifice

image-to-video

stable-video-diffusion
dynami-crafter
streaming-t2v
ms-img2vid
i2vgen-xl




follow this website on neocities