{"slug": "newsbench-expert-grounded-evaluation-of-epistemic-quality-in-ai-news-reporting", "title": "NewsBench: Expert-Grounded Evaluation of Epistemic Quality in AI News Reporting", "summary": "Researchers introduced NewsBench, an expert-grounded benchmark for evaluating the epistemic quality of AI news reporting, detailed in arXiv paper 2610.33240v1. NewsBench scores quality across three dimensions — source quality, factuality and neutrality — using about 3,000 evaluation prompts, roughly 500 of them refreshed monthly for emerging events, plus about 2,500 human expert-annotated items for calibration and validation, with automated AI judges validated against expert-adjudicated gold labels. Applied to four frontier models, NewsBench revealed distinct, non-overlapping failure profiles in sourcing, factuality and neutrality.", "body_md": "arXiv:2610.33240v1 Announce Type: new \nAbstract: As frontier AI systems increasingly serve as sources of political and current-events information, evaluating the epistemic quality of their responses is now especially consequential. Existing evaluations often focus narrowly on ideological alignment, rely on measures not validated for this setting, or depend on reference labels derived from public opinion or non-expert annotation, potentially producing misleading or miscalibrated results. To address these limitations, we introduce NewsBench, an expert-grounded benchmark that reframes evaluation around editorial standards developed by senior practitioners in journalism, policy, and intelligence analysis. NewsBench operationalizes quality across three dimensions, source quality, factuality, and neutrality, using a two-phase pipeline in which domain experts first define and calibrate evaluation guidelines, which are then transferred to automated AI judges validated against expert-adjudicated gold labels. The benchmark includes approximately $3,000$ evaluation prompts, with roughly $500$ refreshed monthly to capture emerging current events, roughly $2,500$ human expert-annotated items for calibration and validation, and a suite of AI judges designed to enable scalable evaluation. To validate our methodology, we report both inter-rater reliability among human experts and AI judge--human agreement across all three dimensions. Finally, we apply NewsBench to four frontier models, revealing distinct and non-overlapping failure profiles in sourcing, factuality, and neutrality.", "url": "https://wpnews.pro/news/newsbench-expert-grounded-evaluation-of-epistemic-quality-in-ai-news-reporting", "canonical_source": "https://arxiv.org/abs/2610.33240", "published_at": "2026-10-27 04:00:00+00:00", "updated_at": "2026-10-27 04:14:32.326528+00:00", "lang": "en", "topics": ["ai-research", "ai-safety", "large-language-models", "artificial-intelligence"], "entities": ["NewsBench", "arXiv"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/newsbench-expert-grounded-evaluation-of-epistemic-quality-in-ai-news-reporting", "markdown": "https://wpnews.pro/news/newsbench-expert-grounded-evaluation-of-epistemic-quality-in-ai-news-reporting.md", "text": "https://wpnews.pro/news/newsbench-expert-grounded-evaluation-of-epistemic-quality-in-ai-news-reporting.txt", "jsonld": "https://wpnews.pro/news/newsbench-expert-grounded-evaluation-of-epistemic-quality-in-ai-news-reporting.jsonld"}}