Tailoring news content: How Scandinavian mediahouses have tested recommender systems

It is well known how large-scale media houses and technology companies have successfully made use of recommender systems, research indicating substantial improvements of click rates and user satisfaction. 

Smaller media houses

It is less understood how smaller media houses are coping with this new technology, how the technology affects their business models, their editorial processes, and their news production in general. 

In this research article, the authors report on the experiences from numerous Scandinavian media houses like online publications of Polaris Media and broadcasters like NRK. They have experimented with various recommender strategies and streamlined their news production to provide personalized news experiences. 

Man and machine

Interestingly, editor governed media houses have found it undesirable to automate the entire recommendation process. In the intersection between man and machine, many media houses look for approaches that combine automatic recommendations with editorial choices.

Fast! Complex! Dynamic!

These characteristics from the news flow vocabulary are well-known to researchers of the media industry. News are constantly renewed, news personalization is intrinsically very complex due to the whole dynamics of the media industry as well as stakeholders’ conflicting goals.

In AI Magazine, the NorwAI authors argue that recommender technologies, in addition to influencing the content and style of news stories and the working environment of journalists, have been part of a profound digital transformation of the whole media industry. 

Recommender systems

News personalization means that the media outlet is adapting its news content and presentation to individual users’ known or inferred preferences. The underlying technological solution is often referred to as a news recommender system, which is normally embedded into a comprehensive media platform that integrates journalistic work and decisions with large-scale information processing capabilities.

In the article the news organizations’ online business models are discussed and their initial interests in recommender systems. The media industry, along with music business, were among the first commercial sectors to be interrupted and transformed by the irresistible force of the internet. 

The various technologies used in news personalization are then presented, followed by an analysis of the news domain and its suitability for algorithmic recommendations. 

Examples

After a discussion of personalization services among medium sized Scandinavian newspapers, attention is turned to two aspects that have received considerable attention in recent years. 

First, user privacy issues and new regulations are gradually forcing the recommender systems to be more transparent and controlled by their users.  

Second, the news organizations may need to consider potential long-term effects of personalized news like filter bubbles, echo chambers and polarization of readers. From a technology point of view, news personalization can be regarded as one of many AI-driven applications that continue to transform the media industry. 

Societal implications

In the last section a broader look is taken at how artificial intelligence affects not only news organizations’ strategies and business models, but also their journalists and the whole news production process. 

Citation:

Gulla, J., Svendsen, R. ., Zhang, L., Stenbom, A., & Frøland, J. (2021). Recommending News in Traditional Media Companies. AI Magazine, 42(3), 55-69. https://doi.org/10.1609/aimag.v42i3.18146