What are news brands doing with GenAI right now? 

Well, if you recently came back from five days in London for the INMA World Congress of News Media and study tour, your head is no doubt buzzing with ideas and opportunities.

Here’s a look at some of the GenAI aspects of the event:

At Thomson Reuters, AI is seen as a tool to unlock the human potential of journalism — and every staff member is expected to be alert to the possibilities.

The news agency has created a GenAI video tool that produces a time-coded transcript, a translation, and a scene list. Clients can now search within the video for key terms to quickly jump to the right spot. It analyses scenes, frame by frame — much faster than humans do — and identifies key people using facial recognition technology. 

Reuters is also working on training algorithms with technology companies and corporations. The company is looking at licensing opportunities but also at training a model so it can create headlines or content in the voice of an external organisation using Reuters content. Reuters data will also be used to validate content in RAG frameworks. 

At New Zealand Media and Entertainment, GenAI holds opportunities for revenue, CEO Michael Boggs said. 

“We make a lot of money from radio traffic reports,” he told conference attendees, pointing out that soon, these traffic reports would be generated by AI, producing 4,500 minutes a week of content. Next up: weather reports. And then, news, using a human-trained voice to which NZME will own the intellectual property. Also on the roadmap: uses in advertising.

The use of GenAI for formulaic, commodity reporting (like traffic or financial data) actually raises the value of human journalism, Sky News Group Executive Chairman David Rhodes pointed out: “Equity prices and weather are devalued. The value of what the bot cannot tell you is that much greater.”

Humans will seek out human-created and human-curated content, Hearst UK CEO Katie Vanneck-Smith said, pointing to Spotify playlists versus Spotify’s algorithmic selection as an example. The marketing team at Hearst has been using AI to drive subscriptions growth for a while, but the real potential lies in using it to increase productivity and efficiency at magazines, she said. 

How much of a game changer is generative AI?

The limiting factor to innovation is no longer money or tech resources — it is news organisations’ ambition and imagination when it comes to using GenAI, said David Caswell, founder of StoryFlow Ltd. 

He cited Cuestion Publica in Colombia, an investigative news organisation of about 15 people that built a database of relationships in Colombian society it queries when news breaks to see how various people in the story are connected, instantly providing contextualisation of breaking news.

Cuestion Publica further fine-tuned a model to adopt its sassy voice. The automation produces tweets for X within 15 minutes, compared with the three hours it used to take journalists earlier. 

It will be increasingly important for news brands to listen to their audiences and “viscerally” understand them because knowing what they care about is going to become more and more valuable, Caswell said. 

“Think about things that you can only do with these tools — whole new things that you could never do manually,” he said.

If you’d like to subscribe to my bi-weekly newsletter, INMA members can do so here.

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