AI for Newsrooms
Real-world AI projects from newsrooms worldwide, curated news about AI in newsrooms, case studies, reports, researches and guides to help journalists and news organizations understand and implement AI.
AI for Newsrooms. News, projects, reports, guides:
Why your AI survival plan should start with a smile
Some 71% of newsrooms are stuck with AI adoption. There is a way out, writes media and strategy consultant, Katya Gorchinskaya.
WAN IFRA
Storylines
Storylines is The Guardian’s first reader-facing AI tool that quietly enhances existing “tag” pages by using a large language model to analyze the headlines of the latest 200 articles on a topic and cluster them into three clear narrative threads, each labeled with an AI-generated subtitle, so readers can more easily grasp the main storylines without the system ever rewriting or summarizing the journalism itself; it’s explicitly framed as a curatorial aid that highlights and organizes human-edited work, with protections like headline-only inputs, extensive senior-editor review, limited rollout to a handful of tag pages, and a built-in “off switch” to minimize hallucinations and safeguard editorial integrity.
audience engagement🇬🇧 · 21 initiatives
Why The Guardian’s first reader-facing AI product isn’t a chatbot
The Guardian has launched its first reader-facing AI product, Storylines, which uses AI to generate related article links on certain topic pages. Unlike a traditional chatbot, Storylines aims to highlight and showcase The Guardian's journalism by threading related articles into narrative storylines. The AI tool is fed only article headlines, not full text, to prevent misrepresentation and hallucinations. A team of 20 senior editors evaluated the AI's output and provided feedback to refine the model. Storylines is currently in limited testing on 10 tag pages and will not be rolled out across The Guardian's entire corpus. The publisher has a "large red button" to turn off the feature if needed. By taking a cautious approach, The Guardian aims to protect its journalism and readers' trust while exploring the potential of AI.
🎧 Digiday
How to Budget for Your Newsroom’s AI Project
When budgeting for an AI project in a newsroom, it's essential to consider the costs beyond token costs and model choices. According to analysis of over 30 newsroom AI projects, labor costs make up around 65% of budgets, while technology and operations account for 20% and 15%, respectively. To manage staffing costs, a "consultant-first" strategy can be effective, where high-level consultants are used to bootstrap before hiring full-time staff. The cost of hiring AI specialists can vary greatly depending on location and availability. For example, CalMatters paid around $120,000-a-year to a software engineering agency to supplement their in-house engineering talent. AI can also help reduce labor costs by automating manual work, such as data entry. Reuters, for instance, used AI to process handwritten logs of temperatures, saving on manual data entry costs. Effective budgeting requires separating costs into building and running the project, making the case for the project, and potential labor cost savings.
Generative AI Newsroom
Process Over Persona
Gina Chua discusses the limitations of using prompts to guide Large Language Models (LLMs) to mimic human-like responses. Instead of relying on persona-based prompts, Chua advocates for a process-oriented approach, where LLMs are trained to follow a specific sequence of steps to achieve a task. She demonstrates this approach by working with Claude, an LLM, to develop a tool that deconstructs and analyzes news stories. The tool breaks down a story into its thesis, facts, assumptions, and analysis, and then evaluates the strengths and weaknesses of the argument. The results show that this process-oriented approach can produce more accurate and informative outputs, and highlights the importance of understanding the complementary advantages of humans and machines in working with AI. This approach enables the development of more sophisticated news tools and products.
📨 Restructured News
An AI Upheaval Is Coming for Media
Journalist Nick Lichtenberg uses AI tools to produce a high volume of stories quickly. He uploads press releases and analyst notes into AI tools, prompting them to generate articles that he can edit and publish. In six months, Lichtenberg produced more stories than any of his colleagues at Fortune did in a year, cranking out seven articles in one day. His approach involves AI playing a leading role in researching and writing stories, a method some view as unconventional in journalism. Lichtenberg's output has been exceptionally high, making him an outlier in the field.
The Wall Street Journal
Weekly Newsletter.
One Friday email that rounds up AI for Newsroom.
Chat with AI For Newsroom
Let's Build the AI Index for Newsrooms
Submit AI initiatives your newsroom is working on, resources, policies, or media content to help journalists and newsrooms worldwide stay ahead in the AI era.