This AIArcman graphic illustrates
the manual labour,
industrial machinery and
digital ages
-ts1673969346.jpg)
This website is about Artificial Intelligence Art, presenting examples of images made by Artificial Intelligence (A.I.). You may think these images are a co-opted mingling of styles, or outright theft, sincere flattery or just lazy copying. Whatever your thoughts about the morality of the works, these are new images, not completely in all cases, but enough to consider them separate but linked to a previous work. These are mostly derivative works.
For some people the term derivative is unfavourable, and to some extent it is. We reserve the greatest accolades to those people who are the most creative, such as those who have initiated great changes in art, who have not been derivative, but wholly creative. But what about the rest of us? You could argue that most art is actually derivative since it tends to imitate or respond to style and fashions in art that others have invented.
You may remember that Andy Warhol's images of Queen Elizabeth, Marilyn Monroe and even Campbell's Soup, were all borrowed images, that were then treated in the industrial process of lithography, and then presented as new works.
AI Art is an art, just like photography, that is produced by a digital process, not by hand.
AI art is currently in its infancy. The process can be extremely hit or miss. One can expect ten or twenty rejected images for every acceptable one. And just like any other art, there are good and bad results. A skillful artist can usually coax better images out of AI than a novice, but it is possible to be lucky and dial in the perfect array of text prompts to create something impressive. Some servers are better than others. Some AI models are better at some subjects than other models. Expect rapid change in the AI landscape, measuring improvements by weeks instead of years. Massive undertakings are in process, and people are both excited and fearful about the prospects. Some artists see themselves out of employment, while others are relishing the increases and rapid visualization of possibilities for their own work. Photography caused the same dilemma in the 19th Century. Photography forced painting to re-invent itself, but photography also became a great tool for artists to use.
Presently, images including the human and animal forms can be very difficult to create, but not impossible. Additional fingers, and disjointed limbs are commonplace. Lettering and words for some reason are usually unintelligible. AI is very good with landscape subjects, including buildings, trees and vegetation, and lighting effects can be exceptionally effective. Science fiction imagery, which allows subject matter to be synthetically created and completely unrelated to earthbound forms appears to be its strongest suit. Symmetry and composition are often very well done. Styles from pencil drawing to ink and water colour cartoons, and from photographic images to painterly images are all possible and effective.
AI art works in several forms. The most common at present is the creation of an image based on a "prompt"as they are called, to describe the desired image in words naming as many pertinent details as possible within the maximum allowable characters. Details may include for instance, the place, weather, time of day, characters, historical events or other narrative. In addition, you may specify a style, or artist to help shape the image.
Some servers allow the use of a starting image which the artist provides, as well as a short text prompt. Some servers allow for parts of that starting image to be erased, which are then replaced by the AI in the resulting AI image. It is also possible to add new extensions to a traditional painting, whereby the AI understands how the lines, colours and textures used in the original work can be extended to create a larger version of a historical painting. It is also possible to change the weather in an image, adding snow, sunlight, cloud or rain to create a new ambiance. And to create a darker or lighter time of day or change of season. These approaches are all used in the examples below.
I
Canadian
Artists'
Work
Re-imagined
with a.I.
as
North Shore Lake Superior
by Lawren Harris 1926
Lone Figure
AIArcman
2023
This image was created by removing the centre of the starter image, and also allowing the AI to expand the image slightly on all four sides.
Laurentian Hills
A.Y. Jackson
1933
Laurentian Hills Revisited
AIArcman
2023
The original image was extended in four directions. The AI introduced the dark hills with clearcuts in the distance.
Cold Country Canada No.1
Arcman
Acrylic on canvas 2015
Cold Country No.1 Revisited
AIArcman 2023
Original image was used as starter, AI reinterpreted the scene, increased view on all sides. Additional detail work was done on the image using Corel Painter to create brush textures.
To Prince Edward Island
Alex Colville 1965
Woman With Binoculars
AIArcman
2023
Only the centre of the original image was retained, with AI providing the figure and the background. Several iterations were necessary to produce this result, with additional work in photoshop, and digital painting in between AI iterations.
The West Wind
Tom Thomson
1917
Arbutus Tree, Vancouver Island
AIArcman
2023
The original image was extended on all four sides, location was shifted to Pacific Coast. Ai retained hills in distance but revised waves, foreshore and cloud patterns.
Market Day, Avignon
Arcman
Digital painting
2019
Market Day, Avignon Revisited (SERIES)
AIArcman
2023
This series shows a variety of stylistic and compositional responses to the starting image.
This series of Cave Painting images was created by using written prompts containing a description and a desired style to AI. These images are composites, derivative of a incredibly large database of imagery, but each image is unique.