Making AI work for journalists

Artificial Intelligence is on the horizon everywhere, and it will also change the world of journalism. The question is how to make it work for journalists, rather than replace them.

There seems to be a general fear of robot journalism, while media organisations experimenting with it consistently emphasise the intention to empower their journalists rather than replace them. Many of the automation solutions currently on the market use some form of artificial intelligence. But that does not mean they can do what human journalists can in terms of creativity (for a great explanation of what AI is and isn’t, also see this interview).

Automation solutions for journalism do produce content more quickly, but they can only do so if there are structured databases available to base the stories on. That’s why robot journalism works well for sports, financial and crime reporting. Journalists will still be needed to develop new sources, explore angles and investigate in-depth stories.

“AI is changing the world, it’s on the horizon everywhere, and even local journalism will have to migrate into that world,” says professor Lars Nyre (University of Bergen).

The University of Bergen is working with local newspapers Sunnhordland, Hordaland and Hallingdølen to better understand the role of technological tools in their work today.

Human-centred AI to support journalist creativity

Artificial Intelligence could support journalism as much as undermine journalists; it will all depend on how the technology is used.

INJECT supports the discovery of story angles beyond journalists’ initial reflexes. It helps identify original stories and potential sources more quickly. Its AI-based search algorithms combine Natural Language Processing, Advanced Information Retrieval and a recommendation system to explore content in novel ways.

Journalist Sindre Thoresen Lønnes (Hallingdølen newspaper) investigated a story on spinal muscular atrophy using INJECT to support the exploration of different angles for his articles:

“I found too much inspiration for my story, so then I found the login feature really helpful. That way I could bookmark articles and use them later. I’ve been writing about this medicine for two or three months now.”

Opening up the black-box

In addition to its focus on exploration, INJECT differs from other solutions in that it addresses the feeling of many that they have no control over the algorithms:

“There is a fear of AI, because it is typically black box. People don’t have control over it; it just automates. Whereas ours is not automation. It’s simply using AI to empower you and up-skill you. I think ours is an AI that you can control and influence and use,” professor Neil Maiden (City, University of London) explains.

To develop INJECT, experienced journalists were interviewed on their strategies for coming up with new stories and original angles. Their expertise has been codified into the algorithms in the form of Creative Sparks.

According to a recent Deloitte report, human-centered design is crucial:

“AI applications must reflect realistic conceptions of user needs and human psychology. Paraphrasing the user-centered design pioneer Don Norman, AI needs to ‘accept human behavior the way it is, not the way we would wish it to be’.”


Want to find out how INJECT leverages human-centred AI to support journalists’ work?
Get in touch!