AI recommendation systems are being used by almost 70% of news publishers, according to Journalism, media, and technology trends and predictions 2023 from the Reuters Institute for the Study of Journalism. As the development of AI technologies picks up speed, work is underway to understand the relationship between algorithmic recommendations and reader personas.
Takeaways
- A recent research report from the Belgian University, Katholieke Universiteit Leuven, focused on the design of ‘personae’ with the aim of improving ‘transparency, diversity and agency’ among news consumers.
- The project developed a group of three personas that could be presented to every reader in place of one-to-one personalization. This avoided the need for people to identify with just one persona, giving them greater agency around the recommendations that they receive. They could choose between:
-
- The challenger: Challenging pre-existing beliefs with alternative viewpoints.
-
- The expert: Deeper detail to maximize knowledge gained.
-
- 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 approachable. A different persona could be chosen at different points in the day to match reading mood and needs.
How can publishers use reader personas?
Although the research conducted by the team at KU Leuven has not been rolled out by a publisher yet, it does spark some interesting discussion points for product and data teams. On the blog for digital solutions provider Twipe, Mathew Lynes lists a series of discussion points for publishers.
- Whilst personalization is a positive step, how can publishers give readers a choice between personalization and editorially curated reading experiences?
- With trust in news being at an all-time low, can publishers use clearly identified reader personas as a way to increase transparency and begin to rebuild ‘strained‘ relationships?
- Can publishers with multiple titles use personas to direct readers away from the content in their chosen title to related publications, encouraging readers to expand their subscriptions?
Established use cases
Publishers are already leaning into specific reader personas to keep audiences engaged with a variety of reading experiences.
- L’Édition du Soir, the evening edition of Ouest-France, is focused on the unwinder persona. It includes videos, games, puzzles and also short stories about the main news.
- The expert persona works well for niche publications. The Athletic focuses reporters and data analysts to focus on specific teams and sports, using the deep-dive approach in their delivery of local sports coverage.
- Dutch publisher NRC uses Twipe’s James recommendation AI. In line with the ideas behind the challenger persona it uses a hybrid algorithm to combine trending news with personalised interests. The approach typically exposes readers to 80% to 90% of the day’s news content, introducing them to a broader range of subjects.
Besides the commercial concerns for improved reader engagement, the researchers at KU Leuven believe that the introduction of transparent reader persona options will reduce the potential harm of algorithmic recommendation systems. They concluded that:
In an increasingly automated future, allowing users more control and including them when designing recommender systems is key… we hope that media organizations take up the challenge towards developing human-centered and responsible algorithmic systems that serve the public good.