A quick thread about my book 'Machine Habitus: Toward a Sociology of Algorithms', just out with @politybooks! #MachineHabitus 1/6 
This book's premise is that the social roots and social consequences of machine learning algorithms cannot be fully understood sociologically through popular notions such as 'bias', or 'filter bubble'... 2/6
In the book I argue that most algorithmic systems can be seen as peculiar participants in society, socialised through human-generated training and feedback data that bear the imprint of unequal social worlds. Culture is the seed transforming algorithms into social agents! 3/6
The machine habitus is a set of cultural dispositions encoded in machine learning systems as the result of this data-driven socialisation process. Like Bourdieu's habitus, MH shapes the behaviour of (artificial) social agents as an opaque, generative cultural mechanism 4/6
Recommender systems, image classifiers, autocomplete algorithms, chatbots: these automated systems silently contribute to the techno-social reproduction of social and symbolic boundaries through myriad recursive interactions with users, between habitus and machine habitus 5/6
The book presents simple examples and reviews critical perspectives about a) the social shaping of machine learning and b) the algorithmic shaping of society. MH also provides social (and data) scientists with a new Bourdieusian framework for unpacking the culture in the code 6/6
• • •
Missing some Tweet in this thread? You can try to
force a refresh
