An artistic Turing Test from the 160s

I found a description of the following experiment at link (an intriguing discussion of AlphaGo's famous "random" or "genius" move 37). These two pictures were presented in 1966 by A. Michael Noll, a Bell Labs engineer, to about one hundred of his colleagues, technical and non-technical. The question: “One of the pictures is of a photograph of a painting by Piet Mondrian while the other is a photograph of a drawing made by an IBM 7094 digital computer. Which of the two do you think was done by a computer?” Only about one third of the participants identified the Mondrian painting correctly. And 60% liked the computer version better (without knowing which was which). The two thirds who misattributed the art thought the less uniform distribution of lines was the product of the artist.

Even though the blog's author suggests that the computer version is more random, I would argue that the positions of the lines are more uniformly random in Mondrian's case and that, therefore, a random generative model for the Mondrian is simpler than one that can generate the computer version. One major issue is that this experiment has an n of 1, which makes it hard to come up with a general theory. But one thing is clear, interestingness lies in the eye of the beholder and not in the end product itself. Creativity is (mis)attributed to an assumed, hidden process.

--

Pictures from link

Left - Piet Mondrian, “Composition in line, second state,” 1916-1917. © Collection Kröller-Müller Museum, Otterlo. Courtesy: Collection Kröller-Müller Museum, Otterlo. Right - A. Michael Noll, “Computer Composition With Lines”, 1964. Created with an IBM 7094 digital computer and a General Dynamics SC-4020 micro-film plotter. Photo: © A. Michael Noll