Attention Welfare after Generative AI: Cognitive Quiet as an Institutional Standard

Generative AI lowers the marginal cost of producing plausible, personalized, and interactive communication while leaving human capacities for interpretation and refusal finite. Existing work identifies important parts of the resulting problem through cognitive liberty, mental privacy, attentional autonomy, digital well-being, manipulation, and platform governance. This article argues that these approaches do not yet supply an integrated standard for judging the institutional conditions under which attention remains usable for self-direction and democratic life. It develops attention welfare as a sufficiency-based framework with three dimensions: attentional security, attentional agency, and attentional ecology. The framework is distinguished by its institutional unit of analysis, its focus on the conditions of reception, and its treatment of attention costs as distributable externalities. A structured synthesis of philosophical, behavioral, human-computer interaction, and legal evidence shows why generative systems intensify attention burdens through scalable production, adaptive personalization, simulated social presence, and continuous feedback. The article then justifies a limited right to cognitive quiet and operationalizes it through a seven-factor reasonableness and proportionality test. Applying that test to the European Union's Digital Services Act, AI Act, and adjacent consumer-protection tools reveals significant but fragmented coverage. The proposed framework does not authorize content-based censorship or prescribe an ideal of contemplative life. It identifies a regulatory floor below which communicative infrastructures unreasonably shift the costs of synthetic abundance onto finite minds.