This project is a collaboration with
Carmen Dusmet Carrasco.
What are algorithms?
Do we understand them and do they understand us?
Can they speak democracy?
Human technology vs. algorithm semantic?
In 1751, Dennis Diderot used an Encyclopaedia, in the form of a book,
with an alphabetical index to organise ‘all human knowledge.’
His attempt to categorise information in such a way came form a desire to democratise knowledge
and make it accessible to other social classes, besides the elite.
Today, Google is the main source that organises human knowledge.
The Index, still exists today as a method to classify information, however,
the medium and the language used to do so are very different.
Algorithms have become the new means in which information is categorised.
The purpose of democratising information still remains.
Information is accessible to a higher percent of society today, however,
as Pierre Levy puts it, we interact with a massive media-library (post-encyclopedia);
thanks to the interface, users are not required to comprehend the method in which information is categorised.
But is it necessary to understand the method in order to fully achieve full democratisation of information?
Would a deeper understanding of this computational process change the way we perceive information? And above all,
does the method have to understand us?
If this algorithm can not understand certain terms like we do,
should the responsibility be taken out of it?
There seems to be a lack of semantics in the algorithmic process that denies it access to terms like ‘democracy’.
However, we are still relying on this method to decide for issues as relevant as what political party should we vote for?
This proves that democracy goes beyond calculation tools,
so why still rely on statistical and automated measurements to deal with it? We have to re-think
how to include the algorithm into a democratic process and above all, we must be critical on how much
of our decision making we want to give up. With Current Language we engage on a research-based project where
Google's Cloud Vision API becomes the base ground to explore the above questions.
We interviewed Holger H. Hoos (Computer Scientists),
Pierre Levy (Computer Science philosopher) and Nikos Voyiatzis (Media Artist),
to discuss our findings.
Thanks to Natalia Jordanova
for the pictures.