Universal Adaptive Reader: A Framework for Generated Reading Environments in the AI Era

Humanity has successively developed writing, books, printing, and the Web as media for preserving and transmitting knowledge. Yet reading environments have remained largely provider-designed, requiring readers to adapt themselves to a given presentation. This paper proposes the Universal Adaptive Reader (UAR) as a framework for Generated Reading Environments: reading environments composed at runtime from Person, Goal, Device, and Content while preserving the source as an invariant reference. The contribution is not the novelty of individual functions such as reflow, question answering, concept maps, terminology support, or read-aloud. Rather, UAR treats the composition of the reading environment itself as the primary design object. The paper positions UAR relative to adaptive hypermedia, adaptive reading systems, and recent AI-enhanced readers; distinguishes an Adaptive Reader, which modifies an existing reading environment, from UAR, which composes the environment itself; and defines a source-preserving framework with a reference architecture and principles for provenance, grounding, traceability, human control, and stable interaction. It further relates Generated Reading Environments to the changing role of textbooks and to Capability Continuity. The resulting framework aims to make AI-assisted reading adaptive without making the source itself unstable.