Contribution short abstract:

In this paper, I approach the links between psychological models and technological design through an in-depth study of current research in machine learning. In particular, I elaborate on Google DeepMind’s interest in idleness and trace it back to the 1990s psychology laboratory.

Contribution long abstract:

In this paper, I approach the links between psychological models and technological design through an in-depth study of machine learning research. I elaborate on Google DeepMind’s interest in implementing idleness in machine learning algorithms and trace it back to the 1990s psychology laboratory, when so-called resting state research drew attention to the processes that occur when volunteers’ brains are supposedly at rest.

Resting state research reversed the reigning experimental paradigm of cognitive neuroscience, put an emphasis on cognitive bandwidth, and linked creativity and pathology to the subject’s ability to control "offline" thought. Against this backdrop, mind wandering has been reconceived as a system-critical mode of information processing and mindfulness is increasingly positioned as a strategy to keep the pathological effects of sustained attention at bay.

Since Google’s machine learning algorithms “don’t even have Christmas off” (Google DeepMind’s CEO Demis Hassabis in an interview with the Guardian), researchers are experimenting with mechanisms that implement through code what happens in human brains while we are slacking or asleep. In my paper, I engage with the situated character of the human that these algorithms implement, provide some insights into how this may affect psychological thinking, and suggest avenues for diversifying what machine learning algorithms could be.