Sunday, November 22, 2026 | Jumada al-akhirah 11, 1448 H
broken clouds
weather
OMAN
26°C / 26°C
EDITOR IN CHIEF- ABDULLAH BIN SALIM AL SHUEILI

Journalism can be more than AI fodder

minus
plus

A healthy information environment depends on supply and demand for quality inputs. To remain competitive and independent in an AI-powered ecosystem, news outlets, publishers, research institutes, and NGOs will need to develop their own skills and tools for establishing where information comes from and whether it can be trusted. Otherwise, they will lose the independence they need to set their agenda in dialogue with audiences and serve the public interest.


In September, unredacted files from The New York Times’ lawsuit against OpenAI and Microsoft revealed what one of Microsoft’s own senior scientists called the use of copyrighted material to train AI models the “largest theft of labour in human history.”


Publishers are beginning to push back. According to NewsGuard, 67 per cent of the most-engaged news sites now block crawlers from accessing archives of their content, and the lawsuits are piling up. But blocking is a defensive position, and it leaves an important question unanswered: What could publishers do with those archives themselves?


A news outlet’s archive is one asset AI companies cannot replicate. In Guatemala and South Africa, media organisations and nonprofits have stopped treating their archives as just something to maintain. They’re building chatbots that answer farmers’ questions, tools that let local readers dig into procurement databases, and systems that sort thousands of reader comments a week.


How much influence media outlets have over information markets depends on their capacity to innovate. The problem is that Big Tech still sets the incentives. Publishers are taking LLM developers to court while trying to track where their content ends up and find ways to charge bots and AI agents for access.


Organisations with deep content archives and a willingness to experiment are already using AI to gain competitive advantages that were previously out of reach. For decades, news organisations have competed on the quality of their journalism and sources. But with information more plentiful than ever, they are looking for ways to build trust and reach new audiences. As the economics of information change rapidly, some are selling real-time data and analysis that clients can act on or tapping into crowd-powered prediction markets.


But where does that leave publishers whose value lies in authorship and community? Around the world, independent publishers are developing AI tools of their own, running models locally and carefully mapping their data architecture. They are also training models only on verifiable proprietary data and exploring community-based, open-source, and commercial uses early in product development.


These investments are creating new revenue streams, from native advertising and sponsorships to in-house AI tools. Food for Mzansi, an agriculture-focused publisher in South Africa, has developed Agri-Pulse 360, an AI ecosystem that provides small-scale commercial farmers with round-the-clock access to practical advice. Drawing on the organisation’s eight-year content library with thousands of articles, the chatbot uses retrieval-augmented generation (RAG) to answer farmers’ questions and will eventually be available through WhatsApp. Agri-Pulse 360 can also be used to generate TikTok reels and track newsletter subscribers’ reading habits.


Food for Mzansi wants to attract advertisers through a native advertising tool it calls “Solution Moments.” It is also seeking pilot sponsorships from major brands for AI-generated technical workflows and aims to automate client promotions within targeted newsletters.


In Guatemala, independent news outlet Ojoconmipisto and data provider Sociedad.info launched Pin Local, an AI tool that draws on government-spending records to produce personalised weekly newsletters. Ojoconmipisto’s journalists do the reporting, and Sociedad.info’s AI searches its public-procurement database for information relevant to readers in each of Guatemala’s 22 administrative regions. The same system could work with data from other countries and allow the organisation to sell customised reports on individual procurement contracts.


In Argentina, fact-checker Chequeado is using AI to sift through the thousands of comments it receives from readers each week. Its tool, Nunchi, uses enterprise models to classify these comments by sentiment and intent and then ranks them by relevance. It was even trained to understand humour and irony. Editors get a Slack brief twice a day with recurring topics and questions that haven’t been covered. They can also use a dashboard to see how audience interest is shifting over time.


In Colombia, Cuestión Pública has built six Model Context Protocol (MCP) servers that connect to 13 public-procurement databases, a national business register, and its own reporting archive. The system lets users explore records covering nine million companies, daily political events, and politically connected individuals in a single conversation. Cuestión Pública also tracks token use live on its site. It previously relied on the speed and usability of a chatbot but is now focusing on generating more value through a tool that supports accountability reporting. The organisation also plans to release a privacy-focused version as a paid service.


What drives these projects is the search for new revenue streams beyond traditional news products. Some publishers are attempting to earn audience loyalty by building direct relationships with their readers, while others treat information as a liquid commodity, selling it to AI firms by the unit. @Project Syndicate, 2026


SHARE ARTICLE
Most Read
AI image
Probation period cannot exceed three months for monthly-paid workers National Day drone shows planned in Muscat, Salalah and Manah Oman, Uzbek sign agreement on new bank, visa exemption, strategic partnership Trucks movement to be restricted on three roads on Friday
FOLLOW US
arrow up
home icon