5 minute read
Swiping the news Kai Landolt
A lot has been said about selective information menus; with our algorithmic driven information landscapes being able to offer us customized news based on clicks and likes, we face the risk of excluding ourselves from information that contradicts our bubbles. How can we design a way out of this?
The question of whether we lock ourselves up in information bubbles has been widely discussed in the media. The overarching message is that this is very harmful to society; it would increase polarized discussions and, worse, erode democratic society.
Advertisement
A significant number of academic studies have been looking into this issue. However, their findings are not uniform. A recent monumental study on manipulation, disinformation and radicalization in media and politics, Network Propaganda, has been looking into more than four million articles from 40,000 news sources. The researchers examined whether and to what extent ‘information bubbles’ influenced the 2016 election campaign and Trump’s first year. The core of their findings, presented in colorful graphs, diagrams and word clouds, is as surprising as it is simple. The prevailing idea that Americans live in separate bubbles because of the polarization of the country, which only permeates information confirming their worldview, is not true. Or rather, it’s half true: the right half. The researchers show convincingly that there is indeed a right-wing media bubble, in which propaganda is endlessly pumped around, but not a left-wing. The left seems to be much more “open” and has more cross connections with the mainstream media than the right. Those addicted to Fox News (30 percent of the population, including Trump) hardly receive any information from other sources. The ‘ecosystem’ of the American media is asymmetrical.
Right-wing media critics will see this as confirmation of their conviction that the mainstream media are distinctly left-wing and not objective or neutral. No wonder! But the researchers’ data reveal something else: the use of established news sources (with classic journalistic techniques such as adversarial hearing or verification) has a dampening effect on left-wing expressions at the margins. While on the isolated right that corrective exchange with mainstream media is lacking.
In the Netherlands, a recent study by Rathenau Instituut (Van Keulen, Korthagen, Diederen en Van Boheemen, 2018), shows that the information landscape in the Netherlands is still rather balanced. The majority of Dutch news consumers have a very diverse media menu. However, the study concludes with a warning message; we need to stay alert towards the ‘bubblization’ of society.
How can we design ourselves out of the bubble? That was the principal question of my design driven research. What if we could introduce people to different bubbles, or make them more aware of their rather monotonous news diet?
For this, I first looked into the algorithms themselves. If they are capable of serving us with customized information, could they then perhaps also service us with things we are less familiar too? The question introduced me to the world of ‘reversed algorithms’: the system is reversed and the algorithms are not serving you information based on your clicks and likes, but rather present you the things that you normally would shy away from.
This resulted in a first prototype – ‘reversing the news’ – that introduced people to a more diverse information menu, as well as different sentiments for the news. Most of the users were rather surprised as the test made them aware that they indeed easily skipped topics like
927 WORDS BY KAI LANDOLT
Kai Landolt graduated from Design Academy Eindhoven and is currently working as a researcher in de Werkplaats on ‘swiping the news’, a study into ways people can be helped to step out of their information bubble.
sports or fashion, although most of the people mentioned that my reverse news system would not be very appealing to them. Testing this first prototype made me understand that not only is the algorithmic logic not very useful when it comes to a balanced news diet, it also made me realise that the fact that this logic can’t be traced makes it very complicated for people to break out of it.
As the aforementioned research clearly shows, a more progressive or conservative lens on the news will give different perspectives on current events. In earlier times, readers were very much aware of this, reading newspapers or watching channels that related to their political taste. Now that many channels are blended and consumers often don’t know the source that served them the information, this is becoming very difficult.
These insights resulted in a better prototype – ‘swiping the news’ – that offers news consumers the option to swipe to the left to get a left-wing point of view to an article, and swipe to the right to get a more conservative lens on the news. By means of sentiment analyses, using various news sources, all topics get a sentiment shading. And based on that shading, they will be moved to the left or right axes.
The prototype needs fine-tuning, but first user tests turn out to be promising. The swiping is appreciated, as it is a very subtle interface that people are already accustomed to. Also, the fact that it is up to the user to open his bubble, is appreciated.
Of course, a service like this will not ‘solve’ the bubble discussion, however it could be used to at least ask people to reflect on their daily diets. And as swiping is an easy thing to do, it might seduce them to do a bit of left- or right-wing snacking.
References
Benkler Y., Faris R. and Roberts H. (2018). Network Propaganda. New York: Oxford University Press.
Keulen, I. van, I. Korthagen, P. Diederen en P. van Boheemen (2018). Digitalisering van het nieuws – Online nieuwsgedrag, desinformatie en personalisatie in Nederland. Den Haag: Rathenau Instituut.