From Invisible Algorithms to Interactive Affordances: Data after the Ideology of Machine Learning

Critiques the shift from transparent ordering (alphabetical, chronological) to opaque machine learning-driven curation of online information. Argues that invisible algorithmic sorting represents an ideological choice rather than a neutral optimization, and explores alternative interactive affordances that could return agency to users through graph-based data presentation.

Machine learning-driven curation replaces explicit ordering logics (alphabetical, chronological) with opaque personalized relevance, severing the relationship between structure and meaning. Proposes data-as-graphs as an alternative paradigm that makes the relationships between items visible and navigable rather than hiding them behind black-box recommendation systems.