As I waded through the images generated by the AI, there were a number of painterly images that appeared.

That leads me nicely onto the other side of my proposal – getting the artist to consider the image on the terms of the computer.

In Series 3, Episode 4 of the IT Crowd, Jen has a revelation: 

This may sound an absurd statement but if we consider the images flowing through the internet, then surely the most common filetype is the JPG.

JPG, or JPEG if you prefer, is an acronym of The Joint Photographic Expert Group, the software company responsible for the standard, which is described as a ‘committee (which) has a long tradition in the creation of still image coding standards’. In relation to the current phase of the internet perhaps the ‘Joint Photographic Expert Group’ could be considered as an Elder of the Internet.

Images are typically large files and so are consequently difficult and slow to transfer. JPGs attempt to ease these difficulties by compressing the image reducing the file size and allowing faster rates of transfer.

It may well be worth considering how the JPG works. 

The algorithm begins by breaking the image into square 8×8 pixel sections. It then attempts to generalise these areas, examining nearby colour, brightness and the like. The heavier the compression, the smaller the file size but the more generalised these blocks. This leads to JPG artefacts where detail has been overly generalised.

The JPG algorithm breaks the image down into these squares, then sends the image. The receiving end reconstructs the squares into the image on the viewers’ screen.

So what if the artist were to follow these steps when examining an image? What would the image resemble?