Typing ‘glitch’ into Google I’m reliably informed there are ‘about 170,000,000 results (0.75 seconds)’ and that Glitch is available on Netflix subscription, Amazon Prime from £1.89, YouTube from £1.99 and Google Play also from £1.99.

Handy to know.

That must mean there’s 226,666,666.666 glitches costing somewhere between £2.52 and £2.65 every second. Amazon Prime is the cheapest, maybe I should go with them.

But does anyone actually need 226,666,666.666 glitches and for that matter, what is a ‘glitch’?

The Error & Power residency asked Naho and myself to explore the relationship between our practice and error, particularly focusing on technological error, or glitches. That must mean that a glitch is when something goes wrong. So is it still a glitch if we want it to occur?

My proposal considers this idea of conjuring glitches, both from the machine and the artist.

So let us for a moment consider a photograph. No longer the result of a series of chemical reactions in a dark room, today an image is only an image because an algorithm says it is.

Let us begin by collecting a series of images from the property search website Zoopla, so we can peer into the homes of the anonymous, gaze upon their possessions, upon their personal spaces. Now let us pass these images to an Artificial Intelligence and ask it to consider what it sees, and then let it imagine something new. 

But these AIs are “idiots”, all they know is brute force of thought, to look for repeating patterns, to bludgeon their way to a conclusion – they cannot think for themselves. What the image contains is of no interest to the machine, that it is data is enough.

The resulting images, while seemingly correct to the machine, are to us riddled with mistakes, what comes to pass is fuzzy, almost dream like.

Now let’s have a bespoke algorithm ask the artist to consider the same image on its terms, to paint what the machine instructs. But we are “idiots”, unable to see the image within the simple instructions it tasks us. The resulting image bears no obvious visual connection to the original.

What we are left with is a meeting of excess and reduction where the machine is asked to imagine what there isn’t and the artist is asked to exclude to the detriment of the image. Neither artist nor machine knows the content of the image. 

Here the possibility of the glitch, the error and the unexpected are introduced, a process that is heightened when both resulting images are combined to a new image, something that adds the excess of the digital image to the reduction of the painterly image to create a new whole.

I hope the time spent during the residency will bring about a series of images combining the AIs digital output with the painted image from the artist.