Pedro Oliveira is a Brazilian researcher and artist, his work focuses on political use of sound and political violence. He considers himself a researcher, who’s research expresses itself through art.
He began his talk by outlying three themes he feels are essential to his research and practice:
1). Material Conditions of listening- how sound becomes ‘sound’. How the body perceives frequencies as sonic and vibrational information.
2). Listening as Relation- Not just grasping knowledge, to relate, to ‘give in and give on’ to the event. Who we are and where we are, and what informs the conditions of this encounter with the sound.
3). Legibility and detectability- What listening ‘produces’ about the body, meaning that sound is inherent to a ‘body’. Listening as an act produces the body.
Germany adopting language/accent recognition software in immigration.
The practice of using language/accent recognition in immigration originated in the 90’s, a recording would be made of someones accent, a team of people would then analyse the recordings to determine whether the individual was from the place that they claimed to be from. This is mostly used in cases of asylum seekers as there is often a lack of documentation. In 2017, Germany replaced the teams of linguists with software. An individual would enter a room, speak into a microphone for two minutes and the software would return the probabilities of the persons geographical origins. Pedro Oliveira found interest in the idea of machine listening revealing a truth about a person, external from their claims about themselves. He began researching this process, and collected all of the information he could. He investigated into how many times this software had been used in asylum cases, he found that in a single year the software was used almost 10,000 times. Pedro took a specific interest into the success rates of the software, as he found the idea of 10,000 people’s quality of life being decided by software very worrying. He points out the lack of accountability in this process, and the difficulty of contesting the results. Utilising a freedom of information act he found that the margin of error for this software was 20%, which means that out of these 10,000 cases almost 2,000 were probably wrong.
Pedro was commissioned by the gutter institute in Brussels to do an art piece based on this piece of research. He analysed the paperwork that asylum seekers get when they arrive in Belgium, its a description of what they need to do, how they should behave and what will be done with the recordings of their voices if a dispute was made. Pedro then hired a semi professional choir, and had the leader of the choir select singers from Belgium with a migratory history. He then had them sing the documents written by the Belgium government in a slightly mocking way, like a machine was reading it stretching out syllables. An element of this piece was to highlight these people’s relationship with Belgium, as they identify as Belgians but Belgium doesn’t look at them the same way.