The Upwording of Speech – Rezo, Youtube, Podcasts and Machine Learning
Track 2: Learning Machines
Date: Day 1 – 14:30
Location: Neufert-Mansion, Gelmeroda
Whenever people talk about so-called artificial intelligences, it is mostly in the context of self-driving cars, facial recognition or deep fakes and other basically optical phenomena. Much less attention is usually paid to another computer revolution based on machine learning: the recognition and processing of oral language. Speech-based assistants such as Siri, Cortada and Alexa are more derided than actually used. These technologies, however, do not merely change the interaction with computers, but the interaction with speech itself. Phenomena such as the video of Youtuber Rezo render visible the current emergence of a new situation of speaking.
The talk will highlight how this new way of speaking comes about and shed light on the media-technological conditions that enable it, using contemporary cases such as videos by Rezo and the Youtuber ContraPoints. It will be shown that what at first glance might look like videos, are in fact formats in which machine learning procedures can automatically transcribe, analyze, and translate oral speech in the cloud. It is an upwording of speech, as it were, since these forms of processing organize a similar accessibility for the spoken word as was previously only the case for the written word. Finally, the talk will – in a collaborative exchange with the participants – explore the political and cultural consequences this new orality might have.