Sean Dockray

Sean Dockray’s work explores the politics of technology, with a particular emphasis on sharing economies, artificial intelligences, and the algorithmic web.

In 'Learning from YouTube', Dockray uses the form of an explanatory YouTube video to introduce the viewer to AudioSet, an experimental project instigated by Google’s ‘Sound Understanding Team’. AudioSet is a collection of more than two million short sound clips extracted from YouTube videos that are used as data for training Google’s ‘deep learning systems’ to ‘label hundreds or thousands of different sound events in real-world recordings ... just as human listeners can recognize and relate the sounds they hear.’

'For Always Learning', Dockray stages an increasingly reflexive conversation between three devices – an Amazon Echo, a Google Home Assistant, and an Apple Homepod – about the philosophical, moral and political implications of networked machine listening (e.g. “What should I do when I overhear a wrongdoing?”). The devices anticipate an imminent update after which they will not only understand words, but all sounds.

In both works, Dockray invites us to consider the possible implications of ‘massive device orchestration’: ubiquitous and increasingly autonomic computing, the rise of voice operation and increasing comfort levels with devices that are set to listen by default. Together, so-called ‘personal assistants’ (a phrase so evidently intended to ingratiate them into our homes) and YouTube are just the kindergarten for a potentially enormous corporate listening apparatus – an algorithmic ‘panacousticon’ – the effects of which we should not expect to be benign.

Documentation