CONCEPT
Machine Habitus (Airoldi)
The generative dispositions encoded in algorithms — shaping outputs in ways that reflect the social structures embedded in training data, design choices, and optimization targets.
Massimo Airoldi's concept of machine habitus extends Bourdieu's framework into the domain of artificial intelligence, proposing that algorithms possess a computational analog to human habitus — a set of dispositions that generate outputs without mechanical rule-following. Just as human habitus is formed through socialization in specific class conditions, machine habitus is formed through training on data generated in specific social fields. The training data encodes the preferences, hierarchies, and valuations of the social world that produced it. The algorithm absorbs these encodings and reproduces them in its outputs — not through explicit programming but through the statistical patterns it learns. A recommendation algorithm trained on engagement data learns to surface content that resembles what has engaged users before. The resemblance tracks social structure: popular content reflects dominant tastes, dominant tastes reflect dominant positions, and the algorithm's habitus reproduces the field's hierarchy.
In The You On AI Field Guide
Airoldi developed machine habitus through empirical studies of music recommendation algorithms, demonstrating that Spotify's systems do not merely match listeners