year: 2016
medium: digital tool, drawing
Processing, digital print

Rounding up these series of generating and interpreting digital drawing, I made a flip book that presents my long-lasting miscommunication with CaptionBot. Although the algorythm is based on a huge image library and seems to be able to recognise a photographic image almost as well as I would, it ends up completely helpless in front of a drawing. The dialogue followed a set of simple rules:

1. I give a code-generated drawing to CaptionBot as a starting point
2. I receive a new title as a response
3. I take the same drawing and draw what CaptionBot told me on top of it
4. We do the circle all over again
5. And again