AI art: what does the future hold?
A pair of scientists published a set NFTs to mark the 48th anniversary Picasso's death. They were based on a recreation a lost work believed to have been attributed to Santiago Rusinol. This artwork was found under Picasso's The Crouching Beggar.
This project is called Oxia Palus. It was started in 2019 by Anthony Bourached, George Cann. Their goal is to restore lost artifacts, promote responsible AI use by creative industries, arts education centres and create the jobs of the future.
They used X-Ray fluorescent images from Rusinol and The Crouching Beggar to reconstruct Picasso's artwork. They created a 3D height map to re-layer paint onto the canvas. This captured the artist's style and texture. They can use spectroscopic imaging and AI to create visible traces of an older painting in an artwork. This is important for the conservation of historical art.
These are not future events, but they do point to the possibility of AI and art converged.
Bourached argued instead of using AI for new creative ideas, they are looking backwards. AI's future is more than a tool for creation. It can also be used to preserve the environment.
AI is constantly evolving and artists create art based on machine learning and algorithms. The next step in AI art would be for computers to recognize emotions. AI will be able to recognize emotions in visual culture, which would allow experts to create computers with more emotional intelligence and thus, more human-like machines.
AI can only form opinions as well as data. It is therefore necessary to provide a lot of information to enable a machine understand emotions in art. This is what a group of researchers from Stanford University and Ecole Polytechnique did.
The team created a ArtEmis dataset, which includes machine learning models that aim to understand the relationship between emotion, language and visual content. This set contains over 80,000 images as well as over 400,000 emotional attributes.
Volunteers were asked to describe the emotion they felt about an artwork in one sentence. The algorithm then categorizes artworks into one of eight emotion categories and then explains why the emotion is important.
ArtEmis has shown great promise. In some cases, captions were created to reflect the abstract concept in the artwork. This goes beyond what is possible with a computer.
ArtEmis is a system that the team behind hopes will be used by artists to help them evaluate their work and to ensure that it has the emotional impact they desire.
An article in The Gradient by Fabien Offit argues that artists could shift their focus from AI's aesthetic exploration to its potential for critical analysis. He believes that AI art will be a catalyst for innovation through its ability to critique itself.
AI art, like other forms of innovation will be evaluated in terms of its utility in the real world. The goal is not aesthetic anymore. It will be critical to open up new opportunities or divergent ideas.
The future normalization of AI art will make machine learning a collection of tools that solves many philosophical problems.