How Artificial Intelligence has changed Art History

Monday - June 12 2023, 02:49 UTC - 1 year ago

Modern AI has made remarkable progress creating images in many diverse styles. By examining the evolution of both AI and art history, it becomes clear that these two have profoundly grown together by learning from large datasets. AI has been used to create art from reference images or blend styles together, as well as generate pieces from datasets of landscapes, celebrity portraits, and food presentations.

Artificial intelligence (AI) has made remarkable progress creating images that are not only breathtaking, but astonishingly diverse in style. Ten years ago, such an achievement would have been deemed unlikely by experts. Today, AI can create images using specific artistic styles, such as Van Gogh's unique approach, with an infinite range of variations.This raises an intriguing question. How can a series of instructions running on a computer produce art that rivals human creativity?Art history and AIExamining the similarities between the evolution of AI and art history may provide some answers .

The history of modern science can be characterized by the development of specific models that represent the world. Traditional models are developed using mathematical equations, physics, and logic. For example, Newton's law describes how gravity works with a very simple formula.However, modern AI has become increasingly reliant on generic models that can learn complex relationships from vast datasets .

They do this without encoding explicit knowledge (so-called artificial neural networks). These new models do not rely on physics or complex mathematical equations but rather are built by layering many small computation units. Taken as a whole, these can learn and reproduce any pattern present in the data.Interestingly, the evolution of art mirrors that of AI in many ways. Art has also undergone a significant transformation, from being rooted in explicit knowledge and classical traditions, to embracing approaches that challenge the boundaries of art itself .

Like AI, art has evolved to incorporate a more organic and intuitive approach that emphasizes the discovery and creation of new forms and styles.The era of building accurate modelsFor decades, constructed AI models relied upon analytical solutions and equations. These were defined by a handful of meaningful parameters crafted by human experts. The primary objective of this early iteration of AI was to adjust those parameters to explain experimental data .

Artists also used models and adapted them to represent what they observed. Their models came from the careful study of anatomy, color and shape. For example, during the Renaissance, da Vinci dedicated himself to studying the human form by performing dissections of humans and animals. He constructed a mental model of what a human body should look like, which was then used to accurately reproduce characters or imagine allegorical religious paintings .

In 2004, anatomy professors Massimo Gulisano and Pietro Bernabei used computers to analyze Michelangelo's David, confirming the extreme anatomical accuracy of the sculpture (except for a missing muscle that upon investigation was due to the stone imperfection).These mental models became more sophisticated during the turn of the 16th century, with artists perfecting the texture and vibrancy of fabrics, water, and light .

The data-driven eraWith advances in computing, new methods emerged allowing AI to recognize sophisticated patterns in large datasets. A watershed moment occurred in 2012 when computer scientists trained a large, deep convolutional neural network to rivial humans in recognizing objects.This was only possible as a result of advances in AI that could use large datasets to learn and construct meaningful models from scratch .

The transition to this data-driven approach has had a lasting impact on contemporary art. AI has been employed to recreate art based on reference images, or used to blend diverse styles through Deep Neural Style Transfer. AI-generated art has also been used to generate new images based on remarkable datasets of landscapes, celebrity portraits, and even edible food presentations.