Michel, Cassandra, Google et les autres, a project by Algolit, generates found poetry using sentences from e-traces, a database containing over 20,000 articles on digital surveillance and big data.
Is it possible for algorithms to read a poetic story in an archive of news articles on digital surveillance, updated daily by the same person for fifteen years? Can algorithms read a similar story in the texts of a database as that which is told by its creator?
Algolit investigates these questions using e-traces, a database maintained by artist Michel Cleempoel. The database contains more than 20,000 news articles on digital surveillance and big data, collected over 15 years and organised with more than 3,000 manually added keywords.
For each poem the installation randomly picks a French or English keyword which is linked to at least 500 sentences of less than 13 words. The poem follows an ABBA rhyme scheme, commonly found in Greek tragedies.