Judging by the number of AI use cases that have popped up over the past few weeks, it looks like the last quarter was jam-packed with experimentation for many news organisations across the world.

Ranging from search bots to AI anchors, they tackle common problems that news consumers face: how to find what they are looking for on our sites, how to access information when they are tight on time, and how to understand what the news is about.

Other applications help news producers overcome challenges such as making better use of archives, getting breaking news out quickly, and expanding coverage.

Here’s a glimpse into what our peers in the news media industry are working on and thinking about. Let me know what you’re thinking about!

Sonali

GenAI to serve readers better

The Australian Broadcasting Corporation has built a bot, ABC Assist, that uses semantic search, an LLM, and RAG to trawl its content so news consumers can quickly find what they are looking for on its Web site. It also provides links to the source material so readers can quickly verify what they are reading, according to a blog post by ABC: 

“Our product was designed to assist users, not do their jobs for them. The promise of ABC Assist wasn’t that you could ask it a question and just unthinkingly copy and paste its response into an e-mail, or a report, or a presentation.

“As our users have told us repeatedly, the value of ABC Assist for them is as a search tool. In that sense, the answers that it provides are simply gateways to information, and users need to remain in control of how this information then gets used.

“So providing users with not just the information they need, but the original context of that information and a way to dig deeper, is key.”

The broadcaster is now working on letting users apply search filters for the source documents or time period that ABC Assist uses.

Germany’s Handelsblatt is also working on a similar search project, according to Editor-in-Chief Sebastian Matthes. It is also using cloned voices to read out articles and plans to use AI for fact checking, optimising headings, generating summaries, illustrations and graphics, recognising trends, and newspaper layout.

Another public broadcaster, NOS in the Netherlands, has launched an “easy” version of its daily news bulletin, aimed at older people, non-native speakers, and those with learning problems or disrupted education. It will have fewer images, simpler sentence structure, easier language and fewer topics, explained at a slower tempo.

Meanwhile, U.S.-based ESPN is using GenAI to create summaries of games played in the Premier Lacrosse League and National Women’s Soccer League.

“The AI-generated recaps aim to enhance coverage of under-served sports, providing fans with content that was previously unavailable. These sports do not currently have game recaps on ESPN digital platforms, and these AI-generated recaps will be a tool to augment existing coverage — not replace it.” 

Each AI-generated summary will be reviewed by a human editor, bylined “ESPN Generative AI Services,” and will also carry a tagline at the bottom that flags that it was created by AI.

Czech Radio is running its fourth season of a podcast featuring short stories called Digital Writer, in which humans and AI collaborate to produce the content. 

“The good thing to remember about this project is especially that the main goal is not necessarily to produce the best short audio stories that we can,” said Anna Vošalíková, chief dramaturg in charge of digital content at Czech Radio. “We have human writers who write fantastic pieces and those are, in terms of quality, far superior to what we do at Digital Writer. However, it is also important to understand that one of the values of the project lies in the technology that is being used and the technology that is being showcased.”

GenAI to reach new audiences, work faster

The Economist in the U.K. is using GenAI to translate text and video on its Espresso app into four languages — German, Spanish, French, and Mandarin — as it offers the product free to students across the world. Its editors, who speak these languages natively, run a final quality check on the translations, and the news brand is now working on automating the entire process. 

Meanwhile, Mexico’s Grupo Fórmula is using three AI news anchors to present five or six news stories a day, each to younger audiences on platforms such as TikTok.

“We have some of the most important news presenters in Mexico. We don’t plan to replace them. The news stories that (AI anchors) present are small stories and very focused towards young people who don’t connect well with the old-style newscast,” said Director of Technology and AI Infrastructure Oswaldo Aguilar Castro.

Turkey’s Hurriyet uses GenAI to monitor data and automatically create articles on earthquakes. Its draws on a public database and uses ChatGPT-4o. The publication hopes to develop further content features, such as a timeline of updates.

The Times of India is using algorithmic distribution to surface archival content on its subscriber-only product, showing sustained growth over 18 months.

“The algorithm is successfully tapping into the archive and getting clicks on older stories: 50% of clicks from personalised push notifications are from stories older than 2 days,” Times Internet Senior Director of Product Ritvvij Parrikh said.

This point “is critical because it demonstrates that tapping into evergreen content is a ‘distribution problem.’”

Thomson Reuters uses GenAI to comb through press releases for news and generate a handful of “snaps,” or alerts, for each. The machine flags which part of the press release it got information for each snap from. An editor looks over the snaps, decides which ones are relevant, and checks the information for accuracy, if needed, before hitting the publish button. 

The tool is 98.6% accurate.

And for anyone interested in modular journalism — presenting news in chunks that can be recombined, rather than the classic inverted pyramid, to drive greater engagement and understanding — there is Reuters’ partner Particle News (h/t Ole Reissmann), which Germany’s Axel Springer recently invested in. 

It uses GPT-4o to present bullet point summaries of key news events, drawn from several sources, and it also lets users ask questions about particular news stories “which are then answered visibly for all and incorporated into the news item,” Reissmann pointed out.

Worthwhile links

• GenAI and monetisation: OpenAI now has more than 1 million paid users for the corporate versions of ChatGPT, but it might need to raise prices drastically to cover its costs.
• GenAI and efficiencies: Lionsgate signs a deal that should save it “millions and millions of dollars” to get a new, custom AI model in exchange for its content library.
• GenAI and writing: Journalists and other writers are employed to improve the quality of chatbot replies. The irony of working for an industry that may well make their craft redundant is not lost on them.
• GenAI and copyright: This technology puts considerable strain on copyright’s two most fundamental legal doctrines.
• GenAI and trust: This is an interesting study that strikes at the heart of a challenge that we are facing in journalism: the question of trust in facts.
• GenAI and electricity/water: Tech companies are talking to the U.S. government about the strain on infrastructure and the environment; it takes 519 millilitres of water to generate a 100-word e-mail. 
• GenAI and forecasting: Apparently, a new bot is better than experienced humans.
• GenAI and liability: Are you aware of all your AI-related liability? This paper, titled “Is Your Use of AI Violating the Law,” explores several fascinating facets of how you may be.
• GenAI and models: OpenAI releases a new model, which is good at some things and not at others.
• GenAI and future models: They are only going to get better.

About this newsletter

Today’s newsletter is written by Sonali Verma, based in Toronto, and lead for the INMA Generative AI Initiative. Sonali will share research, case studies, and thought leadership on the topic of generative AI and how it relates to all areas of news media.

This newsletter is a public face of the Generative AI Initiative by INMA, outlined here. E-mail Sonali at sonali.verma@inma.org or connect with her on INMA’s Slack channel with thoughts, suggestions, and questions.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .