Three modalities of algorithmic forgetting: Over-capture,curated,and preemptive

Abstract

This article focuses on how consumer platforms normalize forgetting by design through three widely integrated features: Microsoft Recall (over-capture), Google Ask Photos (curated retrieval), and Apple Journal (preemptive guidance). It develops a tripartite matrix of algorithmic forgetting from these cases, situating them within a broader forgetting ecology. Drawing on empirically grounded, qualitative materials (official platform communications, technology journalism/expert commentary, and selected Reddit discussions), the analysis critically traces how each feature reshapes what becomes visible, retrievable, and documentable in everyday life. Recall shifts forgetting into filtering and self-management under conditions of saturation. Ask Photos organizes remembering through selective retrieval, a form of operative erasure in which traces remain stored but, when not surfaced, slip out of practical reach. Journal steers inscription in advance through prompts and defaults that privilege certain kinds of experiences over others. Across cases, claims of convenience and wellbeing can mask structured omissions, narrowing the space for reflective and contested engagement with the past. The study offers “modalities of algorithmic forgetting” as an analytic lens to explore consumer “AI memory” features and their socio-political effects.