APC2015: Recap

Time for a short recap of the Amsterdam Privacy Conference in which we participated. We had a lot of fruitful disussions, over lunch, coffee, and dinner, but especially during the panel on news personalisation which we organized:

Neil Thurman pointed to the fact that personalisation of news sites is happening in other ways than some have envisioned: “Legacy news brands’ attempts at personalisation have, for the most part stalled, as they’ve been put off by user disinterest and technical challenges.” He concluded that much news personalisation is happening in social media and at other content aggregators rather than on the news sites themselves..

Neil Richards added the perspective of intellectual privacy and noted that free expression requires intellectual privacy, the “protection from surveillance or interference when we are engaged in the processes of generating ideas – thinking, reading, and speaking with confidantes”. That turned out to be a problem which Maurits Martijn, a journalist working at Dutch journalism startup De Correspondent, is confronted with in his work as well, as it characterizes De Correspondent’s considerations regarding the personalisation of news: On the one hand, as people can subscribe to individual journalists’ feeds, their site features a personalisation element; on the other hand, they are hesitant to engage too much in personalisation (with personal recommendations etc.), as the readers might be sceptical about this.


Judith Möller  spoke about the changing nature of the public sphere in a more personalized world of news. As the audience gets increasingly fragmented, we need a new conceptualization of the common core of news that everybody knows and talks about, but also keep in mind that news users were selective about their news consumption long before the advent of personalized news media.  She also pointed out that targeting news stories to users raise another question to academica: Are targeted news more persuasive and effective?

Lucia d’Acunto (TNO) reported about her experiences when developing algorithms for news providers. According to her user appreciate transparency about news recommendation systems as well as the choice to select between personalised and not personalised news sections. She pointed out that at this stage all news recommendation systems include an element of serendipity, that ensures that filter bubbles can never be absolute.

We also presented some first own resuts, based on a literature review of literature on selective exposure, information cocoons, and filter bubbles: