AARON is a robotic system developed over many years by artist and programmer Harold Cohen (1995) that is able to pick up a paintbrush with its robotic arm and paint on a canvas without assistance. It paints people in a botanical garden by not only copying an existing drawing, but creating as many unique drawings on the subject as necessary. AARON has never seen a person walking in a botanical garden, but he has gained knowledge about postures and plants with the help of rulers. AARON's knowledge and the way he uses it is not comparable to the knowledge we humans have and use, because human knowledge is based on experiencing the world, and humans experience the world with their bodies, their brains, their reproductive organs, things that computers don't have.
But just like humans, AARON has acquired its knowledge cumulatively. For example, once he understands the concept of a leaf cluster, he can use this knowledge whenever he needs it. For AARON, plants exist in terms of their size, the thickness of the branches in relation to their height, the rate at which the branches thin out as they grow, the degree of branching, the angle at which the branches sprout, and so on. The same principles also apply to leaf and cluster formation. By manipulating these factors, AARON is able to create a wide range of plant types and will never draw the same plant twice, even if you draw a series of plants of the same species. In addition, AARON needs to know what the human body is made of, what parts it has and how big they are in relation to each other. He also needs to know how the body parts move and what types of joints there are and how mobile they are. Since a body moves coherently and is not just a collection of independently moving parts, AARON also needs to know how the body's movements are coordinated, e.g. what the body needs to do to maintain balance.
Conceptually, this is not as difficult as it may seem, at least for upright postures with one or two feet on the ground. It's just a matter of keeping the center of gravity over the base and, if necessary, using the arms to regain a balanced posture. He also knows about erasure, so that if the human body is partially covered, for example, only one arm and/or one leg can be seen. However, AARON knows that humans normally have two arms and two legs, so if he doesn't cover something, he always draws two limbs of each. This means that AARON cannot "break" any rules and will never "imagine" the possibility of drawing, for example, people with only one leg or other forms of abstraction. Consequently, AARON's creativity is very limited and anything but human. Nevertheless, his paintings have been exhibited at the Tate Modern in London and the San Francisco Museum of Modern Art. In a sense, then, AARON passes a kind of Turing test of creativity, for his works are of a quality that can be exhibited alongside those of some of the best human artists.
Simon Colton's The Painting Fool (Colton et al. 2015) is much more autonomous than AARON. Although the software does not physically apply paint to a canvas, it digitally simulates numerous styles, from collage to brushstrokes. In Colton's words:
The Painting Fool requires minimal instruction and can create its own concepts by searching for materials online. The software does its own research and scrolls through social media sites. The idea behind it is that this approach allows it to produce art that communicates something to the viewer because it is essentially drawing about human experience, about how we act, feel and discuss on the internet.
In 2009, for example, The Painting Fool created his own interpretation of the war in Afghanistan from a newspaper article. The result is a juxtaposition of Afghan citizens, explosions and graves of war victims.
Other examples of the application of computer creativity to painting and other visual arts are the works of Karl Sims and Jon McCormack. "Reaction Diffusion Media Wall" (Sims 2016) is based on the interactive simulation of chemicals reacting and diffusing to create dynamic patterns that correspond to the reaction-diffusion equations that govern biological morphogenesis. This work is on display at the Museum of Science in Boston. Earlier work by Karl Sims includes the use of advanced computer techniques to interactively develop images with his Genetics Images system (Sims 1994).
Jon McCormack is also exploring how biological processes can be successfully applied in his project "Design After Nature" (McCormack 2014). In another project, Creative Ecosystems, he explores concepts and metaphors of biological ecosystems (McCormack and D'Inverno 2012) as a means of enhancing human creativity in digital art.
There are many more examples from the field of visual arts. The ones mentioned here are not only representative, but in my opinion also the most important contributions in their field.