AARON and Watson — Beauty and Brains
AARON is a painter. It has an evolving, although restricted, creative process — exploring possible objects to paint and ways in which to creatively represent them on canvas. The resulting work has been displayed in galleries and bought by collectors. And although Harold Cohen wrote the code that created AARON, AARON creates wholly independent of its inventor. Each painting it produces is unique to Cohen, full of unplanned details and autonomous creative decisions.
The same is true of IBM’s Watson. When the machine so egregiously answered “Toronto” to a question restricted to U.S. Cities in Final Jeopardy!, the creators had no idea how the machine came to that conclusion. Watson has no capacity to explain its reasoning process — perhaps the widest gap between machine intelligence and human intelligence, as Eugene Wallingford explains.
AARON and Watson’s unpredictability mirrors the mystery of our own intelligence. When I started studying AARON in the late 90s, Cohen’s work was a landmark for people interested in creative computing. I was not only drawn to Cohen’s approach to computer science as an artistic path, I was drawn to AARON’S wabi-sabi approach to creating art. Cohen’s thesis — “What are the minimum conditions under which a set of marks functions as a (representational) image?” — is exercised by AARON.
AARON has grown over the decades and its work has become more sophisticated. Earlier paintings consist of black and white line drawings. Cohen had to establish rules for representing real-world objects. These rules come effortlessly to humans. They’re shared across cultures and they have existed since the earliest cave paintings. Cohen was, in a sense, teaching his prodigy the most basic concepts that we take for granted. By the 1990s, however, AARON could understand and use color effectively. The image that accompanies this essay conveys the remarkable result of Cohen’s work.
Unlike Watson, AARON’s visible act of growing up and adhesion to aesthetic simplicity is a mirror for our edification — our mysterious creative process. Exploration is the future of computing. The exploration of space. The exploration of possible futures. Most importantly, the exploration of ourselves and the invisible workings within.
[Note: This is a continuation of a previous essay on Watson, creativity and artificial intelligence]