[2507.22893] Towards a New Science of AI as Cognitive Infrastructure

View a PDF of the paper titled Towards a New Science of AI as Cognitive Infrastructure, by Giuseppe Riva

View PDF

Summary:Modern human-AI interplay analysis overlooks how AI programs essentially reshape human cognition pre-consciously, a vital blind spot for understanding distributed cognition. This paper introduces “Cognitive Infrastructure Research” (CIS) as a brand new interdisciplinary area to reconceptualize AI as “cognitive infrastructures”: foundational, typically invisible programs conditioning what’s knowable and actionable in digital societies. These semantic infrastructures transport that means, function by means of anticipatory personalization, and exhibit adaptive invisibility, making their affect tough to detect. Critically, they automate “relevance judgment,” shifting the “locus of epistemic company” to non-human programs. By narrative situations spanning particular person (cognitive dependency), collective (democratic deliberation), and societal (governance) scales, we describe how cognitive infrastructures reshape human cognition, public reasoning, and social epistemologies. CIS goals to handle how AI preprocessing reshapes distributed cognition throughout particular person, collective, and cultural scales, requiring unprecedented integration of numerous disciplinary strategies. The framework additionally addresses vital gaps throughout disciplines: cognitive science lacks population-scale preprocessing evaluation capabilities, digital sociology can’t entry particular person cognitive mechanisms, and computational approaches miss cultural transmission dynamics. To attain this aim CIS additionally offers methodological improvements for finding out invisible algorithmic affect: “infrastructure breakdown methodologies”, experimental approaches that reveal cognitive dependencies by systematically withdrawing AI preprocessing after durations of habituation.