Recommender systems – the basic building block of any website today – determine the news we see, the posts we read, the products we buy, the music we listen to. Most of the time the algorithm behind those systems evaluates what it already knows about us in order to recommend something we might be interested in.
In order to study this new emerging form of consciousness I decided to have an interaction with it in the form of an interview.
On the left-hand side you see a photo I sent to Amazon's recommender system.
On the right-hand side you see the response I got.
The Amazon recommender system works from the Amazon's iPhone app. Their Memo service lets you make a picture note of whatever items you find in your surrounding. It then uses artificial intelligence to match your note to the most appropriate product on Amazon's site, which you can buy. Normally used to make photos of books you may need in order to find them online, it can also be used for things that surround you.
If artificial intelligence doesn't match the product, the request is sent to Amazon's online marketplace MechanicalTurk.Com. There each request can be processed for a few cents by a nameless worker (most likely from India or China). Amazon calls this infrastructure "artificial artificial intelligence", referring to the real machine called Mechanical Turk built in the 19th century as a chess computer, which in fact had a real human inside. So when the real artificial intelligence fails in matching your photo to a product, a real person makes a match for you, effectively functioning as a replacement for artificial intelligence.
Our interaction with the machine and the nameless human workforce behind it is documented in chronological order.
It is presented as a durational installation, a 45-minute long lecture-demonstration, and as a book.
By Dmitry Paranyushkin
Full version of the installation (40 mins long): vimeo.com/75911034
Loading more stuff…
Hmm…it looks like things are taking a while to load. Try again?