AI art generators make digital graphics from short lines of text, or ‘prompts’. They can develop a variety of art mediums, including music, collages, digital art, and others. They analyze existing art styles and compositions using neural networks and machine learning to develop new digital artworks. AI art generators may now be used to make logos, flyers, 3D representations, and more. MidJourney, Jasper, DALL-E, BlueWillow, and Leonardo AI are some AI art generators that have gained immense popularity in recent times.

You need not know any programming language or have any hands-on artistic skills yourself. With these tools, anyone can create images (maybe it cannot all be called "art" yet) with a few clicks and within a few seconds. The image below, for instance, was created using the AI art generator, Leonardo AI, by providing the AI model with a brief series of prompts.

The ‘Human’ Contribution to AI-Generated Art

Let us first understand how AI generates art. Data (in this case, photos) is collected by AI art generators and utilized to train the AI using a deep learning model. Imagine how your drawing teacher in school would show you samples of artwork and ask you to recreate them. Machine learning goes through this same process but with millions of images available on the web or those uploaded by the art creator.

Patterns, such as the style of specific pieces of art, are recognized by the ML model. The AI then uses these patterns to generate unique graphics in response to a user's prompts. These prompts are the most crucial part of the process.

The more detailed the prompt you give, the better the results generated by the AI. Iterative processes often also allow you to fine-tune your intended outcome by making several images. Some platforms even allow you to provide ‘negative prompts’, which help the model exclude things that must not appear in the final creation.

Keeping Real Creativity Alive

Any creative process begins with creative thoughts. A photographer, a cinematographer, a choreographer, a painter, or a writer, all go through a similar process. They visualize the final creation in their minds. They make keen observations of the world around them and seek inspiration from nature and earlier creators.

Cognitive Psychologists describe and measure creativity using the concepts of divergent thinking and remote connections. While the AI has the advantage of having access to millions of ‘inspiring’ images, it still fails to make meaningful associations on some occasions. The most common example of the limitations of current AI art generators is the likelihood of creating human figures with six fingers or four toes.

Can we revisit our traditional definitions of creativity? Can we train the younger generation of artists to use AI as a tool to boost their creations? The core training of keen observations, critical evaluation of existing work, and visualization will remain crucial even in the near future. The ways of assessing and judging creativity would change. For instance, an individual generating more complex and intricate prompts leading to a more realistic and flawless image may be termed more creative than one who cannot.

In short, we will have to become more creative about defining, measuring, and executing human creativity if we do not wish to feel obsolete before the competition given by AI.

This article has been authored by Prof. Sairaj Patki, Faculty of Psychology, FLAME University.