There were a bunch of programmed robots doing stuff like making images that did the rounds on the internet around 2008-09 and there was talk of AI and autonomous robots making art around the corner. I used to follow all of the technology alerts for years and that was near the end of that for me anyway. It got boring but I was interested in machines doing stuff and when I was painting I thought how can I prepare for an AI robot future and I did some paintings on walls that were like my answer to that in 2009.

Like I thought I had to be as crazy as possible because that is a pretty human trait. I had to make stuff that machines couldn’t do because they would be too logical. Then closer to the last few years I had a friend in IT who was using AI just for fun. He was like you can tell it to make a chicken with horse legs and he was riffing on that. To me it kind of sounded pretty pointless. How is a monstrosity like that supposed to be creative? I mean the machine did it but it isn’t something an artist would do.

The other reason I started to make images looser and crazier was to avoid references just being straight copies. Like if I had a photo I would use the structure and just make it new again. I did a similar thing in the early 00s as well where I would use references and layer the structure making coded pieces of hybrid graffiti and reference chiaroscuro kind of things. Even when I do a faithful rendition of a sampled image there are things about it that only make sense as paintings. Like if it was real or rendered from the painting there would be big gaps in logic.

I am sampling a lot of AI material now and I feel I have to change it a lot to make it work better, otherwise it is too lifeless. I have to reimagine it and also it has to fit the context. When I see the form which are mostly digitally rendered I am drawn to any subtleties because that is a sign it has potential. If there isn’t anything subtle about it then you have to imagine those subtleties and make it in the painting. Brian Eno was talking on a podcast recently and said that people get excited about AI producing a video of a dog dancing, but what they miss is the dog isn’t dancing well.

There is instead the surprise and acclaim that it did it but no understanding of if it was done well. I see a lot of AI material on buses and in advertising and it is suited to that space of function over form. It is especially mediocre advertising but does the job. These works I am doing now which are only using partial AI references and drawings become interesting as the form takes on decisions that are happening in the real world through an intuitive process. I am not a robot so I want to see how it develops on the day. Just seeing what needs to be done or what can be done as the paint hits the wall.