The last of our interviews from the MA in Interaction Design show features Anuradha Reddy, who was previously featured on the blog with her work for the “NOT THIS” MA show in September. At the MA Interaction Design graduation show, Anuradha exhibited the project “Algo News”:
My project is about how machines can process and understand data in various ways and develop multiple interpretations of it. It explores one particular type of data: news media. The news content is mistranslated by an algorithm, and each news item is converted into about 10 different news stories based on the original story.
For example, if the original news story says: “I’m extremely sorry”, the tenth mistranslation may be “I’m extremely happy”. It seems like the machines may have some sort of thought- it makes you think about what they actually understand from your data, and it starts opening conversations about machine learning.
What kind of program does it use to make the changes?
It uses a translation program that everybody is aware of: Google Translate; it is programmed with the Google Translate API in Python language. The outcomes are automatic mistranslated tweets, and automatic captions for YouTube videos as well, with ten versions of mistranslated captions for each video. The data is also converted into newspapers, and each newspaper has ten versions of mistranslated content. The newspaper is a direct, physical way of interacting with algorithms, and it questions media itself, it questions authorship, content, authenticity. “Is this item newsworthy at all?” That is a big question that opens up.
What kind of changes in meaning did you notice?
The machine makes associations- in one mistranslation “dumb and dumber” was automatically mistranslated as “Jim Carrey”, which brings the realisation that machines don’t think in just one way. Where are they learning this from?
Pop culture, yes, so the machine starts to explain our own aesthetics and our own culture by going through this process of mistranslation, and that was a huge lesson for me to learn.
How was the Twittter feed received?
The Twitter account managed to gather a lot of attention thanks to its use of keywords and it is very often interacted with by other accounts; I get a lot of favourites and retweets, but I have very few followers, as human followers take me for a spammer.
Are you going to continue this project in any way after graduation?
I have several ideas. I may turn it into a website or I may produce weekly issues of content from different news sources. I am also thinking about commercialising it, perhaps through a Kickstarter campaign where something like this gets printed and distributed on the London tube.