Artificial Intelligence is being used by nearly all major news brands. Kasper Lindskow, head of research and innovation at Ekstra Bladet, said the company is among those strategically using data to drive growth. 

Ekstra Bladet is the largest news brand in Denmark, with tabloid-style journalism content focusing on news, sports, and entertainment. The company has been through a massive digital transition over the last decade and continually is transitioning in the digital market.  

In the early 2000s, Ekstra Bladet used analysis of historical data and manual reporting for growth. Around 2015, this practice evolved into proactive use of live data in the organisation. Last year, the company began automating processes and decision making. 

 But Lindskow said this is the time to use Artificial Intelligence. 

“The space is now mature enough for us to develop some basic solutions that are fairly well known, and we are fairly sure they will create value in the business,” he said. 

AI is increasingly important for a competitive news experience for readers. Lindskow believes this will become even more important in the future, specifically for the news industry. 

Ekstra Bladet plans on getting ahead of the curve with AI by developing it within the business now. The goal is not to replace major online platforms, but use the same tools as the major online platforms, such as YouTube. If the company does not take action, Lindskow said, users will find the news elsewhere, and the brand will become obsolete.

To ensure this does not happen at Ekstra Bladet, three teams explore and implement the growth of the brand: a data science team, an innovation team, and a team of partnerships that include researchers, post-doctorates, and a Ph.D.

The ultimate goal for AI at Ekstra Bladet is to contribute relevant and engaging content, while also being informational to readers. 

Ekstra Bladet is mainly a “text-based” news brand, focusing on core areas for a text-driven news brand. The company leverages recommender systems (RS) through computer vision and speech recognition, and natural language processing (NLP) through churn prediction, lead scoring, and survival analysis, as well as reinforcement learning. 

The company publishes 350 news stories a day, but most of these stories are lost in the feed because the headline and front-page stories are the same for all readers, not customised to readers’ interests. Lindskow estimates this leaves about 300 stories unnoticed daily. 

The RS creates opportunities to match users with news they want to see and match niche news stories with specific groups of readers who are interested in that content. 

One of the NLP developments Ekstra Bladet is working on allows users to search for topics or subjects within the article rather than having to do a Google search. 

“Natural language processing is the machine learning method that classifies text, lays text, and generates text,” Lindskow said. “This is very important for making an algorithm understand the text.”

Lindskow gave an example of a more advanced project on how to generate content that involves partnerships with universities. One project is working on abstractive summarisation solutions, which he said some might call “the holy grail.” This solution would allow companies to generate a summary of an article in a new language. The main problems facing this technology are translating factually correct information from the article, and ensuring translations don’t come across as bland and uninteresting to readers.

Ekstra Bladet strives to create short-term and long-term value through AI, but it comes with challenges, such as the speed of news, and how the interests of users are constantly changing. Ekstra Bladet wants these changes to always align with its mission and values, so this is a challenge that must be addressed before launched to users. 

This case study originally appeared in the INMA report, The Guide to Smart Data Strategy in Media.

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