I met Gretchen Andrew when we were both panelists at a MTArt conference on art and artificial intelligence.  I was sharing some of the work Vastari is developing for machine learning, to match content with venues, and as a result also moderating a panel discussion on how artificial intelligence may affect the way we define, appreciate and preserve art in the future. The panel included an architect, an artist and a conservator, which was an amazing debate.

Gretchen, the artist on the panel, was sharing her Internet Imperialism project where she looks at the internet as a much trusted but tenuous form of authority that can be used to understand, manipulate, and imperialize definitions.  Her work considers the content that currently dominates search results and how, for example, Amazon has come to mean a company before a river.

This favoring of, “companies over communities and products over people,” occurs because of how information is structured, something Gretchen considers to be one of the most powerful tools of our time.  And with the internet is being used to train artificially intelligent systems, any blind spots, bias, or structural preferences risk being codified into our future.

Gretchen uses the intersection of her digital art and painting practice to bring awareness to the limitations of how the internet thinks.  We invited her to share her work at our Hoxton square office where we continued the conversation we started in Cambridge.  


Tell us about the work you shared with our office?

These flower paintings are versions of Billy Childish paintings that I learned to make while working in his studio. When I first started making work in his Billy’s style I’d put them online as “after Billy Childish paintings” the traditional method of attribution.  Search engines, which often fail to grasp nuance, started to return them AS Billy Childish Paintings. The search engine is right in one respect but its truth is incomplete.  From this, I started my practice of exposing where search engines fail by replacing existing search results with my paintings. My hope is that the paintings serve as a reminder that the other content returned by the search engine is also influenced by personal perspective as we know paintings to be.


What’s the connection between this flowers series and your Internet Imperialism and #accordingToTheInternet projects?

I always hope my paintings stand on their own merit and so understanding their relationship to my Search Engine Art is optional.  But #accordingToTheInternet these flower paintings are Billy Childish paintings.  I call this Internet Imperialism:  the intentional structuring of information to dominate the online definition or reputation of a place or idea.  I’ve taken this method and aggressively applied it to other series of paintings paired with their own search terms.  For example, I am working on a painting series that will imperialize search results for “Hackney Wick” where a couple of my studios have been knocked down.  


You worked at Google out of University how does this have a bearing on your search engine art?

While I worked at Google for a couple years, these ideas came even more from my degree in Information Systems which taught how businesses could use information to create and maintain a competitive advantage. At the time I was taught to think of this as windows vs apple but applied culturally it is Cherokee Tribe vs the car made by Jeep of the same name.  

How does this connect with the implications of artificial intelligent we were discussing in Cambridge? What do you think the biggest risks of using artificial intelligence are?

Artificially intelligent systems work by looking at existing information and teach themselves to make future classifications based on that information.  When search for “person” the top image results are very narrow in terms of diversity.  But AI is learning to identify what a person is and is not based on this information.  The risk is that we back ourselves into a corner where, even if we understand the limitations, we are still dependent on very flawed systems.