News Personas and the Future of Editorial AI Recommendations

 

A recent research report from the Belgian University of Katholieke Universiteit Leuven, focused on designing reader personas with the goal of improving transparency, diversity, and actionability among news consumers.

Three personas were developed, which each reader could select for himself - i.e. the reader himself decides which persona he most closely corresponds to. And the reader always has this choice. This avoided readers having to identify with only one persona. They were given more freedom of choice in the recommendations they receive. Readers could choose between the following three personas:

1. The challenger: challenging preexisting beliefs with alternative viewpoints.
2. The expert: deepening the details to maximize the increase in knowledge.
3. The unwinder: lighter articles on topics of existing interest.

The team gave each persona a visual identity to increase the transparency of the algorithms and make them accessible. Depending on the reading mood and need, a different persona could be chosen at different times of the day.

How can publishers use reader personas?

The KU Leuven study offers some interesting discussion points for product and data teams in editorial offices or publishing houses.

On the blog "Twipe" Mathew Lynes lists the following points:

• How can publishers give readers a choice between personalization and editorially curated reading experiences?
• Can publishers, whose trust in news is at an all-time low, use clearly identified reader personas to increase transparency and restore "strained" relationships?
• Can publishers with multiple titles use personas to direct readers from their chosen title's content to related publications, encouraging readers to increase their subscriptions?

Beyond the commercial considerations of improving reader engagement, KU Leuven researchers believe that introducing transparent reader persona options will reduce the potential harm of algorithmic recommender systems. They concluded that: In an increasingly automated future, it is crucial to involve users in the design of recommender systems. Hopefully, media organizations will rise to the challenge of developing human-centered and responsible algorithmic systems that serve the public good.