📨 From today's weekly newsletter that we do every Friday (subscribe here):
I’ve read Web4.0 manifest this week.
I still don’t believe in a future where AI fully creates, reads, and engages with content for itself while humans simply watch from the side. What feels more realistic for me is the emergence of two parallel content systems:
• a human system, where content is created for people, with context, intent, and real value
• an AI system, where content is processed, synthesized, and used to make machines more useful for us
In that sense, maybe what we are seeing right now is a kind of universal sandbox — a big experimental layer filled with low-value content that AI systems learn from and that will eventually be cleaned up. Because honestly, a lot of AI-generated content today still feels like a fun game to play.
We’re all playing with the tools, generating texts, posts, summaries, images, and endless derivatives. But I suspect we need to go through this phase fully to understand its limits. At some point, the industry (and what is more important — the audience) will realize that volume without value is just noise. And right now the amount of this noise is enormous. In many spaces it already overshadows content that actually matters.
Maybe I’m too optimistic (or even too romantic), but I still believe that eventually some kind of “overall editor” layer will emerge. Whether it’s human curation, algorithmic filtering, or a mix of both, low-quality content will be filtered out because it simply creates no value. In the long run, I still believe the value-first approach wins.
—
Sergei Yakupov
I’ve read Web4.0 manifest this week.
I still don’t believe in a future where AI fully creates, reads, and engages with content for itself while humans simply watch from the side. What feels more realistic for me is the emergence of two parallel content systems:
• a human system, where content is created for people, with context, intent, and real value
• an AI system, where content is processed, synthesized, and used to make machines more useful for us
In that sense, maybe what we are seeing right now is a kind of universal sandbox — a big experimental layer filled with low-value content that AI systems learn from and that will eventually be cleaned up. Because honestly, a lot of AI-generated content today still feels like a fun game to play.
We’re all playing with the tools, generating texts, posts, summaries, images, and endless derivatives. But I suspect we need to go through this phase fully to understand its limits. At some point, the industry (and what is more important — the audience) will realize that volume without value is just noise. And right now the amount of this noise is enormous. In many spaces it already overshadows content that actually matters.
Maybe I’m too optimistic (or even too romantic), but I still believe that eventually some kind of “overall editor” layer will emerge. Whether it’s human curation, algorithmic filtering, or a mix of both, low-quality content will be filtered out because it simply creates no value. In the long run, I still believe the value-first approach wins.
—
Sergei Yakupov
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ Smart Search
by Handelsblatt 🇩🇪
Handelsblatt built a conversational search tool on a unified content database that answers reader questions directly from verified sources while refusing to answer when lacking sufficient information, integrating cross-promotion recommendations to boost subscriber engagement and retention across its ecosystem.
→ Read more on wan-ifra.org
▸ Sakal Media Group AI OCR for Ads
by Sakal Media Group 🇮🇳
Sakal Media Group deployed an AI-powered OCR system to automatically recognize and structure advertising content, cutting manual data entry workload in sales operations and enabling faster workflow processing across its regional publications.
→ Read more on wan-ifra.org
──────────
💡 RESOURCES
──────────
▸ OSINT Navigator
OSINT Navigator is a free, searchable database of open-source intelligence tools organized by use case and data type, designed to help journalists quickly identify relevant research resources for investigations. The platform includes descriptions, URLs, and categorization to streamline the process of finding tools for tasks like geolocation verification, financial record searches, and social media analysis.
→ Read more on navigator.indicator.media
──────────
📰 TO READ
──────────
🗞️ Study Finds A Third of New Websites are AI-Generated
A new study shows that approximately 33% of newly created websites are now generated using AI tools, raising questions about content quality and authenticity across the web. The finding underscores how rapidly AI-powered site builders are becoming mainstream for startups and small businesses, with implications for how newsrooms evaluate online sources and verify information.
→ Read more on 404media.co
🗞️ Hearst rebuilds around data and AI
Hearst is reorganizing its publishing operations to prioritize data infrastructure and AI capabilities as core business functions rather than supporting roles. The restructuring aims to improve targeting, personalization, and revenue optimization across its portfolio of magazines and digital properties.
→ Read more on Tomorrow's Publisher
📨 Signals at Scale: A Look Into the Agentic Future of News
An analysis of how AI agents could autonomously monitor information signals across the web and synthesize them into newsworthy narratives, potentially reshaping how newsrooms identify and develop stories at scale. The piece examines the technical feasibility and editorial implications of agentic systems that operate with minimal human intervention in the news discovery pipeline.
→ Read more on Substack
👀 🇸🇪 We say we want to “build relationships” – but we recruit for production
Swedish newsrooms claim relationship-building is central to their strategy but hire primarily for production capacity, creating a structural mismatch between stated goals and actual resource allocation. The analysis suggests this gap undermines long-term audience engagement efforts when staffing decisions prioritize immediate output over community development.
→ Read more on Dagens Media
──────────
🤖 INITIATIVES
──────────
▸ Smart Search
by Handelsblatt 🇩🇪
Handelsblatt built a conversational search tool on a unified content database that answers reader questions directly from verified sources while refusing to answer when lacking sufficient information, integrating cross-promotion recommendations to boost subscriber engagement and retention across its ecosystem.
→ Read more on wan-ifra.org
▸ Sakal Media Group AI OCR for Ads
by Sakal Media Group 🇮🇳
Sakal Media Group deployed an AI-powered OCR system to automatically recognize and structure advertising content, cutting manual data entry workload in sales operations and enabling faster workflow processing across its regional publications.
→ Read more on wan-ifra.org
──────────
💡 RESOURCES
──────────
▸ OSINT Navigator
OSINT Navigator is a free, searchable database of open-source intelligence tools organized by use case and data type, designed to help journalists quickly identify relevant research resources for investigations. The platform includes descriptions, URLs, and categorization to streamline the process of finding tools for tasks like geolocation verification, financial record searches, and social media analysis.
→ Read more on navigator.indicator.media
──────────
📰 TO READ
──────────
🗞️ Study Finds A Third of New Websites are AI-Generated
A new study shows that approximately 33% of newly created websites are now generated using AI tools, raising questions about content quality and authenticity across the web. The finding underscores how rapidly AI-powered site builders are becoming mainstream for startups and small businesses, with implications for how newsrooms evaluate online sources and verify information.
→ Read more on 404media.co
🗞️ Hearst rebuilds around data and AI
Hearst is reorganizing its publishing operations to prioritize data infrastructure and AI capabilities as core business functions rather than supporting roles. The restructuring aims to improve targeting, personalization, and revenue optimization across its portfolio of magazines and digital properties.
→ Read more on Tomorrow's Publisher
📨 Signals at Scale: A Look Into the Agentic Future of News
An analysis of how AI agents could autonomously monitor information signals across the web and synthesize them into newsworthy narratives, potentially reshaping how newsrooms identify and develop stories at scale. The piece examines the technical feasibility and editorial implications of agentic systems that operate with minimal human intervention in the news discovery pipeline.
→ Read more on Substack
👀 🇸🇪 We say we want to “build relationships” – but we recruit for production
Swedish newsrooms claim relationship-building is central to their strategy but hire primarily for production capacity, creating a structural mismatch between stated goals and actual resource allocation. The analysis suggests this gap undermines long-term audience engagement efforts when staffing decisions prioritize immediate output over community development.
→ Read more on Dagens Media
💡 Quote of a Day:
Aversion to AI writing is rooted in the following two truths:
1. Writing is thinking made visible.
2. Publishing your writing is a vulnerable act.
When someone clearly used AI to write something, they are trying to bypass an inherent part of the public writing process – the revelation of one's inner world.
They believe that the AI writing will act as a buffer and keep them safe from rejection.
But this bypassing breaks trust with the reader, who was expecting to witness and experience the poster’s courage to be cognitively exposed.
The result is disgust and judgment by the audience – the exact reactions they were trying to avoid in the first place.
—
Victoria Gamlen on LinkedIn
Aversion to AI writing is rooted in the following two truths:
1. Writing is thinking made visible.
2. Publishing your writing is a vulnerable act.
When someone clearly used AI to write something, they are trying to bypass an inherent part of the public writing process – the revelation of one's inner world.
They believe that the AI writing will act as a buffer and keep them safe from rejection.
But this bypassing breaks trust with the reader, who was expecting to witness and experience the poster’s courage to be cognitively exposed.
The result is disgust and judgment by the audience – the exact reactions they were trying to avoid in the first place.
—
Victoria Gamlen on LinkedIn
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ Infographic Creator
by Deccan Herald 🇮🇳
Deccan Herald's CMS-embedded AI tool automatically converts finished articles into structured infographics in under a minute, cutting manual production time by roughly 90% while letting editors retain full control over accuracy and tone. The system works across English and Kannada content and generates derivative formats for social platforms, addressing the practical constraint of time-poor editorial teams.
→ Read more on wan-ifra.org
──────────
💡 RESOURCES
──────────
▸ AppliedXL
AppliedXL's research examines which organizations are positioned to profit from AI-powered fact-checking and verification tools, analyzing the competitive landscape between tech platforms, news publishers, and specialized startups. The analysis identifies business models and market barriers that will determine which players can build sustainable revenue from truthfulness infrastructure.
→ Read more on appliedxl.com
──────────
📰 TO READ
──────────
👀 Google On Publishing Commodity Content
Google's search ranking system penalizes low-value commodity content that lacks originality or unique insight, affecting how news outlets compete for visibility. Publishers relying on wire-service feeds or generic reporting risk lower rankings unless they add substantial editorial analysis or reporting.
→ Read more on SE Roundtable
──────────
🤖 INITIATIVES
──────────
▸ Infographic Creator
by Deccan Herald 🇮🇳
Deccan Herald's CMS-embedded AI tool automatically converts finished articles into structured infographics in under a minute, cutting manual production time by roughly 90% while letting editors retain full control over accuracy and tone. The system works across English and Kannada content and generates derivative formats for social platforms, addressing the practical constraint of time-poor editorial teams.
→ Read more on wan-ifra.org
──────────
💡 RESOURCES
──────────
▸ AppliedXL
AppliedXL's research examines which organizations are positioned to profit from AI-powered fact-checking and verification tools, analyzing the competitive landscape between tech platforms, news publishers, and specialized startups. The analysis identifies business models and market barriers that will determine which players can build sustainable revenue from truthfulness infrastructure.
→ Read more on appliedxl.com
──────────
📰 TO READ
──────────
👀 Google On Publishing Commodity Content
Google's search ranking system penalizes low-value commodity content that lacks originality or unique insight, affecting how news outlets compete for visibility. Publishers relying on wire-service feeds or generic reporting risk lower rankings unless they add substantial editorial analysis or reporting.
→ Read more on SE Roundtable
We have small update on our website.
The list of sources (websites and newsletters) that we read to stay updated about AI in newsrooms implementations.
Feel free to add your website, blog, podcasts or newsletter, using 'Add resource' button on the top right of the page.
The list of sources (websites and newsletters) that we read to stay updated about AI in newsrooms implementations.
Feel free to add your website, blog, podcasts or newsletter, using 'Add resource' button on the top right of the page.
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
💡 RESOURCES
──────────
▸ GitHub
Talkie is an open-source language model framework designed for conversational AI applications, offering newsrooms a customizable foundation for building voice-based or text-based interactive tools without relying on proprietary platforms. The project provides modular components for dialogue management and integration with various backend systems, making it relevant for newsrooms exploring chatbot or voice assistant applications.
The catch: it was trained on data BEFORE 1930s.
→ Read more on github.com
▸ Center for Journalism & Liberty
A new report maps how AI companies are negotiating licensing deals with publishers, revealing a concentrated market where a few major tech firms control access to training data and the terms increasingly favor large media organizations over smaller outlets.
→ Read more on static1.squarespace.com
──────────
🫡 AI POLICIES
──────────
▸ How LAist uses Artificial Intelligence
LAist 🇺🇸
LAist has developed a transparent AI policy covering editorial uses including transcription, research assistance, and headline generation, while explicitly prohibiting AI use for reporting, source replacement, and content creation without human review and byline attribution. The station's framework prioritizes human journalists in decision-making roles and requires disclosure when AI tools assist in story production.
→ Read more on laist.com
──────────
📰 TO READ
──────────
👀 How to reduce your token footprint
A guide to minimizing token consumption when using language models, covering techniques like prompt engineering, caching strategies, and model selection to lower API costs and computational overhead in newsroom AI workflows.
→ Read more on Olereissmann.com
🗞️ 🇩🇪 How vibe-coded websites put passwords and health data at risk
German websites increasingly rely on AI-generated code from services like Supabase, creating security vulnerabilities that expose user passwords and health data to unauthorized access. The practice, termed "vibe-coding," prioritizes speed over security audits, leaving newsrooms and other organizations exposed to data breaches.
→ Read more on Zeit
👀 🇫🇮 Shared memory, open models, automation – the new infrastructure of journalism is being built now, and it's not just a tech thing
Newsrooms are establishing shared knowledge systems, adopting open-source AI models, and implementing automation workflows as foundational infrastructure—changes that require organizational and editorial decisions, not just technical ones.
→ Read more on Numeroidentakaa
──────────
💡 RESOURCES
──────────
▸ GitHub
Talkie is an open-source language model framework designed for conversational AI applications, offering newsrooms a customizable foundation for building voice-based or text-based interactive tools without relying on proprietary platforms. The project provides modular components for dialogue management and integration with various backend systems, making it relevant for newsrooms exploring chatbot or voice assistant applications.
The catch: it was trained on data BEFORE 1930s.
→ Read more on github.com
▸ Center for Journalism & Liberty
A new report maps how AI companies are negotiating licensing deals with publishers, revealing a concentrated market where a few major tech firms control access to training data and the terms increasingly favor large media organizations over smaller outlets.
→ Read more on static1.squarespace.com
──────────
🫡 AI POLICIES
──────────
▸ How LAist uses Artificial Intelligence
LAist 🇺🇸
LAist has developed a transparent AI policy covering editorial uses including transcription, research assistance, and headline generation, while explicitly prohibiting AI use for reporting, source replacement, and content creation without human review and byline attribution. The station's framework prioritizes human journalists in decision-making roles and requires disclosure when AI tools assist in story production.
→ Read more on laist.com
──────────
📰 TO READ
──────────
👀 How to reduce your token footprint
A guide to minimizing token consumption when using language models, covering techniques like prompt engineering, caching strategies, and model selection to lower API costs and computational overhead in newsroom AI workflows.
→ Read more on Olereissmann.com
🗞️ 🇩🇪 How vibe-coded websites put passwords and health data at risk
German websites increasingly rely on AI-generated code from services like Supabase, creating security vulnerabilities that expose user passwords and health data to unauthorized access. The practice, termed "vibe-coding," prioritizes speed over security audits, leaving newsrooms and other organizations exposed to data breaches.
→ Read more on Zeit
👀 🇫🇮 Shared memory, open models, automation – the new infrastructure of journalism is being built now, and it's not just a tech thing
Newsrooms are establishing shared knowledge systems, adopting open-source AI models, and implementing automation workflows as foundational infrastructure—changes that require organizational and editorial decisions, not just technical ones.
→ Read more on Numeroidentakaa
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ Guardian Ask the Archive (Internal)
by The Guardian 🇬🇧
The Guardian created an internal AI chatbot that queries the publisher's archive to surface and summarize past coverage, streamlining background research for reporters working on stories across any topic. The tool is particularly valuable for accessing older articles that predate modern search optimization, making it faster to identify relevant historical context and reporting.
→ Read more on niemanlab.org
──────────
💡 RESOURCES
──────────
▸ GitHub Pages
Researchers studied the impact of AI-generated text on the internet from 2022 to 2025. By mid-2025, 35% of new websites were classified as AI-generated or AI-assisted, up from zero before ChatGPT's launch in 2022. The study found a decrease in semantic diversity and an increase in positive sentiment with more AI-generated text.
→ Read more on ai-on-the-internet.github.io
──────────
📰 TO READ
──────────
🗞️ US publishers back Amazon in AI agent access dispute with Perplexity
Major US publishers are supporting Amazon's position in a dispute with Perplexity over AI agent access, signaling industry alignment on controlling how their content is used by competing AI platforms. The disagreement centers on terms for allowing AI systems to access publisher content for training and real-time search results.
→ Read more on PressGazette
📨 🇸🇪 News, collaborations and AI insights
Bonnier News shares updates on organizational developments, partnership activities, and observations about AI applications relevant to Swedish media operations.
→ Read more on Bonnier News
🗞️ How Casey Newton is revamping his newsletter to compete with AI
Casey Newton is shifting his newsletter strategy away from aggregation and analysis toward original reporting and scoops to differentiate from AI-generated content summaries. The move reflects a broader industry challenge: newsletters built on curating existing news face commoditization as AI tools can replicate that value at scale.
→ Read more on NiemanLab
──────────
🤖 INITIATIVES
──────────
▸ Guardian Ask the Archive (Internal)
by The Guardian 🇬🇧
The Guardian created an internal AI chatbot that queries the publisher's archive to surface and summarize past coverage, streamlining background research for reporters working on stories across any topic. The tool is particularly valuable for accessing older articles that predate modern search optimization, making it faster to identify relevant historical context and reporting.
→ Read more on niemanlab.org
──────────
💡 RESOURCES
──────────
▸ GitHub Pages
Researchers studied the impact of AI-generated text on the internet from 2022 to 2025. By mid-2025, 35% of new websites were classified as AI-generated or AI-assisted, up from zero before ChatGPT's launch in 2022. The study found a decrease in semantic diversity and an increase in positive sentiment with more AI-generated text.
→ Read more on ai-on-the-internet.github.io
──────────
📰 TO READ
──────────
🗞️ US publishers back Amazon in AI agent access dispute with Perplexity
Major US publishers are supporting Amazon's position in a dispute with Perplexity over AI agent access, signaling industry alignment on controlling how their content is used by competing AI platforms. The disagreement centers on terms for allowing AI systems to access publisher content for training and real-time search results.
→ Read more on PressGazette
📨 🇸🇪 News, collaborations and AI insights
Bonnier News shares updates on organizational developments, partnership activities, and observations about AI applications relevant to Swedish media operations.
→ Read more on Bonnier News
🗞️ How Casey Newton is revamping his newsletter to compete with AI
Casey Newton is shifting his newsletter strategy away from aggregation and analysis toward original reporting and scoops to differentiate from AI-generated content summaries. The move reflects a broader industry challenge: newsletters built on curating existing news face commoditization as AI tools can replicate that value at scale.
→ Read more on NiemanLab
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
📰 TO READ
──────────
👀 AI-Powered Content Optimization: Tools and Techniques for 2026
Ring Publishing outlines AI tools and methods designed to optimize content performance in 2026, focusing on practical applications for newsrooms seeking to improve audience engagement and editorial efficiency. The guide covers specific techniques for leveraging AI to refine headlines, structure, and distribution strategies based on reader behavior and platform algorithms.
→ Read more on Ring Publishing
📺 The News SEO Playbook with Clara Soteras
Clara Soteras walks through search optimization strategies tailored to newsrooms, covering how to structure content and metadata to improve organic discovery. The discussion covers practical SEO tactics that balance audience reach with editorial integrity for publishers of various sizes.
→ Read more on The Fix
👀 🇫🇮 Chinese AI-bots raided Helsingin Sanomat
Chinese AI scraping bots have targeted Helsingin Sanomat's website, underscoring growing concerns about unauthorized data harvesting for large language model training without publisher consent or compensation. The incident highlights vulnerabilities in content protection strategies as news organizations face increasing pressure from international actors seeking training data.
→ Read more on Helsingin Sanomat
👀 When AI becomes the interface, journalism has to become useful
As AI assistants become primary information interfaces, newsrooms must shift from prioritizing traffic and engagement to delivering genuinely useful insights that AI systems can reliably extract and surface to users. Publishers risk irrelevance if their journalism remains optimized for human browsing rather than machine readability and factual accuracy.
→ Read more on LinkedIn
──────────
📰 TO READ
──────────
👀 AI-Powered Content Optimization: Tools and Techniques for 2026
Ring Publishing outlines AI tools and methods designed to optimize content performance in 2026, focusing on practical applications for newsrooms seeking to improve audience engagement and editorial efficiency. The guide covers specific techniques for leveraging AI to refine headlines, structure, and distribution strategies based on reader behavior and platform algorithms.
→ Read more on Ring Publishing
📺 The News SEO Playbook with Clara Soteras
Clara Soteras walks through search optimization strategies tailored to newsrooms, covering how to structure content and metadata to improve organic discovery. The discussion covers practical SEO tactics that balance audience reach with editorial integrity for publishers of various sizes.
→ Read more on The Fix
👀 🇫🇮 Chinese AI-bots raided Helsingin Sanomat
Chinese AI scraping bots have targeted Helsingin Sanomat's website, underscoring growing concerns about unauthorized data harvesting for large language model training without publisher consent or compensation. The incident highlights vulnerabilities in content protection strategies as news organizations face increasing pressure from international actors seeking training data.
→ Read more on Helsingin Sanomat
👀 When AI becomes the interface, journalism has to become useful
As AI assistants become primary information interfaces, newsrooms must shift from prioritizing traffic and engagement to delivering genuinely useful insights that AI systems can reliably extract and surface to users. Publishers risk irrelevance if their journalism remains optimized for human browsing rather than machine readability and factual accuracy.
→ Read more on LinkedIn
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ Framing Gaza
by Anmat 🇪🇬
A data-driven analysis of 25,000+ English-language news articles from major US, UK, French, and German outlets applies Association Rule Mining to identify patterns in how media outlets frame the Gaza conflict through word co-occurrence in headlines and article bodies. The project offers newsrooms a replicable methodology for detecting framing patterns across outlets and languages using automated article collection and algorithmic analysis.
→ Read more on aifornewsroom.in
▸ The EqualVoice Assistant
by Ringier AG 🇨🇭
Ringier's EqualVoice Assistant flags biased language and stereotypical framing in real time, offering alternative phrasings based on analysis of over 32,000 monthly articles. The tool is now available beyond newsrooms for use in corporate communications, marketing, education, and other fields seeking to reduce representational bias in their writing.
→ Read more on aifornewsroom.in
──────────
📰 TO READ
──────────
👀 Why liquid content is harder than it looks
Adapting content across multiple formats and platforms requires more coordination, technical infrastructure, and editorial oversight than many newsrooms anticipate, particularly when managing quality and consistency at scale. Understanding these operational challenges can help media organizations set realistic timelines and resource allocation for liquid content strategies.
→ Read more on The Media Copilot
👀 We’re making big calls on a technology still in its infancy
Media organizations are making consequential decisions about AI adoption while the technology remains early-stage and its long-term impacts uncertain, raising questions about whether current choices will prove misguided as the field evolves.
→ Read more on LinkedIn
──────────
🤖 INITIATIVES
──────────
▸ Framing Gaza
by Anmat 🇪🇬
A data-driven analysis of 25,000+ English-language news articles from major US, UK, French, and German outlets applies Association Rule Mining to identify patterns in how media outlets frame the Gaza conflict through word co-occurrence in headlines and article bodies. The project offers newsrooms a replicable methodology for detecting framing patterns across outlets and languages using automated article collection and algorithmic analysis.
→ Read more on aifornewsroom.in
▸ The EqualVoice Assistant
by Ringier AG 🇨🇭
Ringier's EqualVoice Assistant flags biased language and stereotypical framing in real time, offering alternative phrasings based on analysis of over 32,000 monthly articles. The tool is now available beyond newsrooms for use in corporate communications, marketing, education, and other fields seeking to reduce representational bias in their writing.
→ Read more on aifornewsroom.in
──────────
📰 TO READ
──────────
👀 Why liquid content is harder than it looks
Adapting content across multiple formats and platforms requires more coordination, technical infrastructure, and editorial oversight than many newsrooms anticipate, particularly when managing quality and consistency at scale. Understanding these operational challenges can help media organizations set realistic timelines and resource allocation for liquid content strategies.
→ Read more on The Media Copilot
👀 We’re making big calls on a technology still in its infancy
Media organizations are making consequential decisions about AI adoption while the technology remains early-stage and its long-term impacts uncertain, raising questions about whether current choices will prove misguided as the field evolves.
→ Read more on LinkedIn
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
📰 TO READ
──────────
🗞️ OpenAI launches self-serve ad platform
OpenAI has launched a self-serve advertising platform, allowing businesses to run ads directly without requiring a dedicated account manager. The move mirrors standard practices at Google and Meta, potentially lowering barriers for smaller advertisers to reach audiences through OpenAI's properties.
→ Read more on Axios
👀 What we heard in Perugia about AI rewriting the rules of journalism
Journalists and media leaders at the Perugia International Journalism Festival discussed how AI is fundamentally changing reporting practices, from newsgathering to distribution, with particular focus on editorial workflows and ethical guardrails. Key takeaways centered on the need for newsroom training, transparent AI use policies, and maintaining human judgment in story selection and verification.
→ Read more on The Fix
🗞️ Major publishers sue Meta for copyright infringement over AI training
Major publishers are suing Meta for using their copyrighted content to train AI models without permission or compensation. The lawsuit represents one of the first coordinated legal challenges from legacy media against tech companies' AI data practices.
→ Read more on Guardian
🗞️ Thomson Reuters Leans on Legal AI Services to Battle SaaS-pocalypse
Thomson Reuters is shifting focus toward AI-powered legal services to offset declining revenue from traditional software subscriptions as customers reduce their spending. The strategy reflects how legacy media and information companies are repositioning around AI capabilities to maintain growth amid SaaS market saturation.
→ Read more on A Media Operator
👀 Turn your data into decisions: 3 things your business needs for growth in the AI era
Google outlines three foundational elements—data integration, AI-powered analytics, and actionable insights—that businesses need to leverage data effectively for decision-making in an AI-driven environment. The guidance emphasizes moving beyond data collection to practical application, particularly for marketing and commerce teams seeking competitive advantage.
→ Read more on Google
──────────
📰 TO READ
──────────
🗞️ OpenAI launches self-serve ad platform
OpenAI has launched a self-serve advertising platform, allowing businesses to run ads directly without requiring a dedicated account manager. The move mirrors standard practices at Google and Meta, potentially lowering barriers for smaller advertisers to reach audiences through OpenAI's properties.
→ Read more on Axios
👀 What we heard in Perugia about AI rewriting the rules of journalism
Journalists and media leaders at the Perugia International Journalism Festival discussed how AI is fundamentally changing reporting practices, from newsgathering to distribution, with particular focus on editorial workflows and ethical guardrails. Key takeaways centered on the need for newsroom training, transparent AI use policies, and maintaining human judgment in story selection and verification.
→ Read more on The Fix
🗞️ Major publishers sue Meta for copyright infringement over AI training
Major publishers are suing Meta for using their copyrighted content to train AI models without permission or compensation. The lawsuit represents one of the first coordinated legal challenges from legacy media against tech companies' AI data practices.
→ Read more on Guardian
🗞️ Thomson Reuters Leans on Legal AI Services to Battle SaaS-pocalypse
Thomson Reuters is shifting focus toward AI-powered legal services to offset declining revenue from traditional software subscriptions as customers reduce their spending. The strategy reflects how legacy media and information companies are repositioning around AI capabilities to maintain growth amid SaaS market saturation.
→ Read more on A Media Operator
👀 Turn your data into decisions: 3 things your business needs for growth in the AI era
Google outlines three foundational elements—data integration, AI-powered analytics, and actionable insights—that businesses need to leverage data effectively for decision-making in an AI-driven environment. The guidance emphasizes moving beyond data collection to practical application, particularly for marketing and commerce teams seeking competitive advantage.
→ Read more on Google
All 287 initiatives from aifornewsroom.in now have some analysis.
Read the latest AI implementation report based on what we've collected so far.
Read the latest AI implementation report based on what we've collected so far.
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ D/Cipher+
by People Inc 🇺🇸
People Inc has launched D/Cipher+, an AI-powered ad targeting tool designed to generate advertising revenue independent of traditional website traffic and reduce reliance on Google-dependent income. The product signals the company's broader push to diversify its revenue model beyond conventional digital advertising channels.
→ Read more on pressgazette.co.uk
▸ Radio France + NotebookLM
by Radio France 🇫🇷
Radio France used NotebookLM to synthesize simultaneous broadcasts from 44 local stations into a structured regional breakdown of farmer concerns in under an hour, enabling same-day briefing for a Prime Minister interview during an agricultural crisis. The workflow—capturing audio via Yacast, processing through a engineered prompt, and extracting quotes and regional data—required minimal technical infrastructure but demonstrated how AI can convert distributed reporting into actionable intelligence for national coverage.
→ Read more on journalismai.info
▸ IMAGINE
by IGN Entertainment 🇺🇸
IGN Entertainment's IMAGINE platform combines behavioral analytics with independent research data to generate audience insights reports that track media consumption patterns across generational cohorts, helping media buyers understand engagement differences by age group.
→ Read more on variety.com
---
→ Check all 287 AI implementations
──────────
🫡 AI POLICIES
──────────
▸ Trust & Safety
Classical KING FM 98.1 🇺🇸
Classical KING FM 98.1 outlines its approach to content moderation and user safety across digital platforms, addressing how the station evaluates AI-generated or AI-assisted content in its programming and community spaces. The policy clarifies standards for removing harmful content and the role of human oversight in AI deployment decisions.
→ Read more on classicalking.org
──────────
📰 TO READ
──────────
🗞️ Google highlights links from subscribed publications in new AI Overviews update
Google's latest AI Overviews update visibly distinguishes links from news publications users subscribe to, potentially directing traffic to paywall-protected content and rewarding outlets with paid subscriber bases. The change signals Google's effort to balance AI-generated summaries with meaningful traffic referrals to publishers, though the practical impact on subscription conversions remains unclear.
→ Read more on Nieman Lab
🗞️ AI-generated 'Epstein Files' podcast hits 2 million downloads, raising alarms over invisible editorial judgment
An AI-generated podcast about the Epstein case reached 2 million downloads without disclosure of its synthetic nature, prompting concerns about how algorithmic curation and generation can obscure editorial decisions that human audiences typically expect to see. The case illustrates a gap between AI capabilities and existing newsroom accountability standards around attribution and transparency.
→ Read more on The Media Copilot
──────────
🤖 INITIATIVES
──────────
▸ D/Cipher+
by People Inc 🇺🇸
People Inc has launched D/Cipher+, an AI-powered ad targeting tool designed to generate advertising revenue independent of traditional website traffic and reduce reliance on Google-dependent income. The product signals the company's broader push to diversify its revenue model beyond conventional digital advertising channels.
→ Read more on pressgazette.co.uk
▸ Radio France + NotebookLM
by Radio France 🇫🇷
Radio France used NotebookLM to synthesize simultaneous broadcasts from 44 local stations into a structured regional breakdown of farmer concerns in under an hour, enabling same-day briefing for a Prime Minister interview during an agricultural crisis. The workflow—capturing audio via Yacast, processing through a engineered prompt, and extracting quotes and regional data—required minimal technical infrastructure but demonstrated how AI can convert distributed reporting into actionable intelligence for national coverage.
→ Read more on journalismai.info
▸ IMAGINE
by IGN Entertainment 🇺🇸
IGN Entertainment's IMAGINE platform combines behavioral analytics with independent research data to generate audience insights reports that track media consumption patterns across generational cohorts, helping media buyers understand engagement differences by age group.
→ Read more on variety.com
---
→ Check all 287 AI implementations
──────────
🫡 AI POLICIES
──────────
▸ Trust & Safety
Classical KING FM 98.1 🇺🇸
Classical KING FM 98.1 outlines its approach to content moderation and user safety across digital platforms, addressing how the station evaluates AI-generated or AI-assisted content in its programming and community spaces. The policy clarifies standards for removing harmful content and the role of human oversight in AI deployment decisions.
→ Read more on classicalking.org
──────────
📰 TO READ
──────────
🗞️ Google highlights links from subscribed publications in new AI Overviews update
Google's latest AI Overviews update visibly distinguishes links from news publications users subscribe to, potentially directing traffic to paywall-protected content and rewarding outlets with paid subscriber bases. The change signals Google's effort to balance AI-generated summaries with meaningful traffic referrals to publishers, though the practical impact on subscription conversions remains unclear.
→ Read more on Nieman Lab
🗞️ AI-generated 'Epstein Files' podcast hits 2 million downloads, raising alarms over invisible editorial judgment
An AI-generated podcast about the Epstein case reached 2 million downloads without disclosure of its synthetic nature, prompting concerns about how algorithmic curation and generation can obscure editorial decisions that human audiences typically expect to see. The case illustrates a gap between AI capabilities and existing newsroom accountability standards around attribution and transparency.
→ Read more on The Media Copilot
🛠️ New AI tools for newsroom
LiquidAmber Discovery
LiquidAmber Discover is an AI-powered research platform for local newsrooms. It helps journalists unlock the value hidden in decades of archived reporting, turning newsroom history into a searchable, actionable knowledge base. The platform supports investigative and enterprise reporting, surfaces story leads from public documents and meeting records, and helps reporters work with the depth and speed of a larger research team. Built for local media, LiquidAmber Discover is designed not to replace journalists, but to strengthen reporting using the full weight of a newsroom’s own institutional memory.
---
Script-to-video
(The Telegraph)
Turn a narration script + slides into a polished narrated video with text-to-speech, transitions, and annotations.
---
Buried Signals
(Spotlight)
Runtime-agnostic OSINT investigation system for journalists. Verified findings, independent fact-checking, knowledge vault ingestion — driven by any agent harness that can read AGENTS.md and dispatch 13 abstract verbs.
---
→ Check all 22 AI tools for newsroom
LiquidAmber Discovery
LiquidAmber Discover is an AI-powered research platform for local newsrooms. It helps journalists unlock the value hidden in decades of archived reporting, turning newsroom history into a searchable, actionable knowledge base. The platform supports investigative and enterprise reporting, surfaces story leads from public documents and meeting records, and helps reporters work with the depth and speed of a larger research team. Built for local media, LiquidAmber Discover is designed not to replace journalists, but to strengthen reporting using the full weight of a newsroom’s own institutional memory.
---
Script-to-video
(The Telegraph)
Turn a narration script + slides into a polished narrated video with text-to-speech, transitions, and annotations.
---
Buried Signals
(Spotlight)
Runtime-agnostic OSINT investigation system for journalists. Verified findings, independent fact-checking, knowledge vault ingestion — driven by any agent harness that can read AGENTS.md and dispatch 13 abstract verbs.
---
→ Check all 22 AI tools for newsroom
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ Semafor Intelligence
by Semafor 🇺🇸
Semafor built an AI system that converts hundreds of event transcripts into vector embeddings to automatically identify recurring themes and key quotes, which journalists then verify and shape into curated briefings with source links. The tool transforms large volumes of spoken content into manageable strategic summaries while maintaining human editorial control over the final output.
→ Read more on niemanlab.org
---
→ Check all 288 AI implementations
──────────
📰 TO READ
──────────
📨 Local journalism doesn't have to be a charity case
Local news outlets can build sustainable business models by focusing on reader revenue and community engagement rather than relying primarily on grants and donations. The piece examines practical strategies for monetizing local journalism through digital subscriptions, memberships, and direct audience support.
→ Read more on News Machines
──────────
🤖 INITIATIVES
──────────
▸ Semafor Intelligence
by Semafor 🇺🇸
Semafor built an AI system that converts hundreds of event transcripts into vector embeddings to automatically identify recurring themes and key quotes, which journalists then verify and shape into curated briefings with source links. The tool transforms large volumes of spoken content into manageable strategic summaries while maintaining human editorial control over the final output.
→ Read more on niemanlab.org
---
→ Check all 288 AI implementations
──────────
📰 TO READ
──────────
📨 Local journalism doesn't have to be a charity case
Local news outlets can build sustainable business models by focusing on reader revenue and community engagement rather than relying primarily on grants and donations. The piece examines practical strategies for monetizing local journalism through digital subscriptions, memberships, and direct audience support.
→ Read more on News Machines
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
🤖 INITIATIVES
──────────
▸ dpa-iq
by dpa 🇩🇪
dpa-iq is an API-based content platform that lets newsrooms integrate verified dpa articles, images, video, and structured data directly into AI agents and automated workflows, enabling applications like auto-generated briefings without manual wire searching. The system works with popular tools like OpenAI and n8n, includes granular access controls, and provides rights-cleared material from dpa and partner datasets.
→ Read more on wan-ifra.org
---
→ Check all 290 AI implementations
──────────
📰 TO READ
──────────
🗞️ Mediahuis joins ‘NATO for news’ AI licensing group SPUR
Belgian media group Mediahuis has joined SPUR, a licensing consortium that negotiates collective AI training rights on behalf of news publishers, adding to growing industry momentum around coordinated licensing rather than individual deals. The move reflects publishers' strategy to standardize compensation terms as generative AI companies seek to license news content at scale.
→ Read more on Press Gazette
──────────
🤖 INITIATIVES
──────────
▸ dpa-iq
by dpa 🇩🇪
dpa-iq is an API-based content platform that lets newsrooms integrate verified dpa articles, images, video, and structured data directly into AI agents and automated workflows, enabling applications like auto-generated briefings without manual wire searching. The system works with popular tools like OpenAI and n8n, includes granular access controls, and provides rights-cleared material from dpa and partner datasets.
→ Read more on wan-ifra.org
---
→ Check all 290 AI implementations
──────────
📰 TO READ
──────────
🗞️ Mediahuis joins ‘NATO for news’ AI licensing group SPUR
Belgian media group Mediahuis has joined SPUR, a licensing consortium that negotiates collective AI training rights on behalf of news publishers, adding to growing industry momentum around coordinated licensing rather than individual deals. The move reflects publishers' strategy to standardize compensation terms as generative AI companies seek to license news content at scale.
→ Read more on Press Gazette
🛠️ New AI tool for newsroom
Spec-Kit
(Github)
Spec-Kit is Github toolkit that turns AI coding into a spec-driven workflow with plans, tasks, and review checkpoints before code starts.
---
→ Check all 23 AI tools for newsroom
Spec-Kit
(Github)
Spec-Kit is Github toolkit that turns AI coding into a spec-driven workflow with plans, tasks, and review checkpoints before code starts.
---
→ Check all 23 AI tools for newsroom
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
📰 TO READ
──────────
📨 Talking about ‘AI’ – a call for more nuanced terminology
Journalists should distinguish between different AI technologies—generative models, machine learning, and narrow automation—rather than using "AI" as a catch-all term, since the capabilities and limitations vary significantly across applications. Using precise terminology helps audiences better understand what a system actually does and avoids overstating capabilities or risks associated with a specific technology.
→ Read more on Substack
🗞️ Thomson Reuters boss says AI licensing deals only involve archive text
Thomson Reuters' chief executive clarified that the news organization's AI licensing agreements restrict training data to archived content, excluding current reporting or real-time news. The statement addresses publisher concerns about how their content is being used to train artificial intelligence systems.
→ Read more on Press Gazette
🗞️ Google launches AI training programme aimed at smaller publishers
Google is launching a training program designed to help smaller publishers understand and adopt AI tools in their newsrooms. The initiative addresses a gap where larger media outlets have more resources to experiment with AI while smaller operations struggle to keep pace with the technology.
→ Read more on Hold The Front Page
──────────
📰 TO READ
──────────
📨 Talking about ‘AI’ – a call for more nuanced terminology
Journalists should distinguish between different AI technologies—generative models, machine learning, and narrow automation—rather than using "AI" as a catch-all term, since the capabilities and limitations vary significantly across applications. Using precise terminology helps audiences better understand what a system actually does and avoids overstating capabilities or risks associated with a specific technology.
→ Read more on Substack
🗞️ Thomson Reuters boss says AI licensing deals only involve archive text
Thomson Reuters' chief executive clarified that the news organization's AI licensing agreements restrict training data to archived content, excluding current reporting or real-time news. The statement addresses publisher concerns about how their content is being used to train artificial intelligence systems.
→ Read more on Press Gazette
🗞️ Google launches AI training programme aimed at smaller publishers
Google is launching a training program designed to help smaller publishers understand and adopt AI tools in their newsrooms. The initiative addresses a gap where larger media outlets have more resources to experiment with AI while smaller operations struggle to keep pace with the technology.
→ Read more on Hold The Front Page
🛠️ New AI tool for newsroom
Academic Research Skills
(Github)
Academic Research Skills (ARS) is a Claude Code plugin that turns AI into a rigorous, human-in-the-loop copilot for the whole academic workflow—from deep literature research and PRISMA-style systematic reviews, through planning and drafting papers with style calibration and citation checking, to multi‑agent peer review, integrity verification, and final formatting in APA/IEEE/other styles.
---
→ Check all 24 AI tools for newsroom
Academic Research Skills
(Github)
Academic Research Skills (ARS) is a Claude Code plugin that turns AI into a rigorous, human-in-the-loop copilot for the whole academic workflow—from deep literature research and PRISMA-style systematic reviews, through planning and drafting papers with style calibration and citation checking, to multi‑agent peer review, integrity verification, and final formatting in APA/IEEE/other styles.
---
→ Check all 24 AI tools for newsroom
⚡️ Daily Round Up of what is interesting in AI for Newsrooms today:
──────────
💡 RESOURCES
──────────
▸ Google Slides
I don't have the ability to access external URLs or view the contents of Google Slides presentations. To provide an accurate preview, I would need you to share the actual text, title, or description of the presentation content.
Could you paste the title and key details from the Google Slides resource?
→ Read more on docs.google.com
──────────
📰 TO READ
──────────
👀 OpenAI launches the OpenAI Deployment Company to help businesses build around intelligence
OpenAI has created a new unit called the OpenAI Deployment Company to help organizations implement AI systems into their operations. The move signals OpenAI's shift toward providing infrastructure and consulting services alongside its model development, targeting enterprises looking to integrate AI into existing workflows.
→ Read more on OpenAI
🗞️ Prolific finance journalists facing questions over identities
Several prolific finance journalists have drawn scrutiny over whether their identities are authentic or AI-generated, raising questions about credential verification in specialized reporting. The incidents highlight gaps in newsroom vetting processes and the growing difficulty of distinguishing between human and synthetic bylines in digital publishing.
→ Read more on Press Gazette
──────────
💡 RESOURCES
──────────
▸ Google Slides
I don't have the ability to access external URLs or view the contents of Google Slides presentations. To provide an accurate preview, I would need you to share the actual text, title, or description of the presentation content.
Could you paste the title and key details from the Google Slides resource?
→ Read more on docs.google.com
──────────
📰 TO READ
──────────
👀 OpenAI launches the OpenAI Deployment Company to help businesses build around intelligence
OpenAI has created a new unit called the OpenAI Deployment Company to help organizations implement AI systems into their operations. The move signals OpenAI's shift toward providing infrastructure and consulting services alongside its model development, targeting enterprises looking to integrate AI into existing workflows.
→ Read more on OpenAI
🗞️ Prolific finance journalists facing questions over identities
Several prolific finance journalists have drawn scrutiny over whether their identities are authentic or AI-generated, raising questions about credential verification in specialized reporting. The incidents highlight gaps in newsroom vetting processes and the growing difficulty of distinguishing between human and synthetic bylines in digital publishing.
→ Read more on Press Gazette
👀 Another big update on AIForNewsroom.in. We already have several tools for news media: ROI Calculator, LLM Tracker and AI Policy Generator. And now we have one more: AI Visibility Checker.
We have to be prepared for the new reality of AI search (you know, AEO, GEO, Google Zero etc). So now you can check if your website is AI ready:
👉 https://aifornewsroom.in/ai_visibility
robots.txt, sitemap visibility, structured metadata, homepage crawlability and AI citation readiness — check all these in seconds and even get recommendations.
Also check the reports we've done, using this tool: https://aifornewsroom.in/ai_visibility/reports
We have to be prepared for the new reality of AI search (you know, AEO, GEO, Google Zero etc). So now you can check if your website is AI ready:
👉 https://aifornewsroom.in/ai_visibility
robots.txt, sitemap visibility, structured metadata, homepage crawlability and AI citation readiness — check all these in seconds and even get recommendations.
Also check the reports we've done, using this tool: https://aifornewsroom.in/ai_visibility/reports
AI For Newsrooms
AI Visibility Check — Free AI Search Audit
Audit robots.txt, schema, sitemap, and article HTML for AI crawlability and citation readiness.