Piece I
My second graffiti piece entirely generated by artificial intelligence (AI), i.e. machine learning (Mars, March 2020). First, an image was generated by training a machine learning model (StyleGAN) on ~2500 frames of a filmed sequence of images of a blue and white „Ajax“ piece in NYC in 1973 (Source: @tarajiqueen).

The output by the machine learning model was then used as an inspiration and practical sketch for painting a new piece that is inspired by the original work and an homage to the early NYC graffiti pioneers of the 1970s. This piece is the second in a series of new generated GAN-generated graffiti pieces or what I call "GANfitti".

The paintings also served as an exploration of Augmented Intelligence Reality (AIR), combining AI graffiti painting with Augmented Reality (AR). The talented Jonas Johannson programmed custom AR filters that could be used on a phone with Instagram and thus the painting could be viewed as a 3-dimensional landscape. I especially liked the idea that any painting can easily become a sculpture this way and that movement and distortion can induce another feel to the flat 2d image. In this case, Jonas came up with a breathing pattern, which went really well with the windy desert backdrop.