Criticisms of generative and database art.

A very interesting talk at EVA 2017 from Jon McCormack. He discussed three criticisms of generative art:

1. ‘algorithmic genericism’: ‘works made with similar generative procedures – even by different artists – possess a certain generic and repetitive character.’ He used the Voronoi algorithm as an example of a technique which is becoming too familiar. (I’m still working through Generative Design, which quotes a works by Golan Levin, from 2000, using Voronoi. Shame, because it’s a nice effect and I hadn’t got that far through the book yet.)

2. works are not ‘true to process’ because they ‘do not exhibit the characteristics of emergent outcomes that the artists claim the processes used represent’. Discussing this, he distinguishes between two forms of autonomy. The first is when the impersonal components of the system self-generate results. The second involves the human ideas of intention, desire and belief. If it possessed the second form, any creative software would need to make the choice to make art, not be explicitly programmed to do so.’ (A challenge here, I think?)

3. Can a machine be ‘authentic’ enough to produce art? He argues that ‘a programme that can change itself in response to external stimulus has the potential to learn and adapt, hence it can do things that the programmer never anticipated…’

McCormack’s comments made me think also about what you might call ‘database art’. A surprising number of speakers at EVA had built art works around downloading some data stream and visualising it, in ways that varied from the elaborate to the quite simple. Two of them even cited the same stream – data about meteorites hitting the earth. I think that this sort of art is open to the same criticism as ‘a generic and repetitive character’.

Also, sometimes there doesn’t seem to be any relationship between the data and the final work. For instance, Scott Kildall’s visualisation of meteorite impact data as holes cut into stone, works, but I feel that his ‘Data Crystals’, 3D printed sculptures of open data from San Francisco about crime and the location of parking meters, fail. You look at an elaborate shape and you are told that this represents homicides in SF or whatever, but it might just as well represent penguins having sex in Antarctica or the tea breaks of chartered accountants in Twickenham. Neither the source nor the expression illuminates the other. (And also he used a clustering algorithm to make his data take a particular shape, in effect destroying it in the process.)

Despite this I have blogged separately a list of data sources available via the net.

I’ve also used an image produced by my own ‘The Portrait Machine’ as an illustration, just to break the blog page up a bit. This is a ‘generative’picture of me. I hope it doesn’t commit any of the sins discussed in this post.

Leave a Reply

Your email address will not be published. Required fields are marked *