If you watched the Netflix series “Black Mirror” then you are probably familiar with the series’ sceptical view on the future that is presented in various dystopia over the episodes. If you have also watched Black Mirror’s movie named “Bandersnatch” from 2018, then you also know that it is already possible that the same content (i.e. the same movie) is presented quite differently to the people that watch it, depending on their preferences and choices during the movie (The Verge, 2019).
From an information management perspective, this new ‘personalisation’ trend is however both interesting and alarming, because more drastic forms of personalisation (i.e. “over-personalisation”) can be applied to more serious contexts as well, for example in the news industry.
Why should we care?
According to Nielsen Norman Group (2019), the over-personalisation of content can lead to “content fatigue” as well as a lack of diversity in what people read on a daily basis. On the one hand, content fatigue means that over-personalisation leads to a more homogeneous experiences for users over time, which is bad for news companies because consumers can easily lose interest in reading more content, after which they can cancel their subscriptions. On the other hand however, the lack of diversity in news content can lead to consumers that increasingly only read (i) information that they already know, and (ii) opinions with which they already agree. This phenomenon, in which people will increasingly find and read content that is in line with their prior knowledge, is known as the “confirmation bias” (Nickerson, 1998).
Furthermore, almost a decade ago, Thurman and Schifferes (2012) already found that a growing number of news websites use “implicit” compared to “explicit” personalisation (i.e. mainly inferring news preferences from the collection of data instead of user input), without the users even knowing about it. Given the duration of this trend, and the possibilities that big data and Artificial Intelligence bring to build detailed consumer profiles, the question rises where this growing act of personalisation will ever stop.
We are almost all familiar with seeing better ‘related’ news articles, depending on our interests and previously read items, which is fine and useful. But what if over-personalisation will go one step further and change the actual content of a news article itself, based on your interests and search behaviour? According to Bodó et al. (2019) and Chrysanthou et al. (2020), different news companies are already experimenting with this form of personalisation.
A personal prediction.
Personally, I think that the risk exists that news companies will over-personalise news content, to the extent that one can generally not anymore have a decent discussion with friends about a specific news article of the same company, because the content of that same article is just too different between you and your friends.
This brings a discussion to life, regarding the extent to which these news companies should be allowed to change the actual context of an article, given the large amounts of implicit knowledge that these companies have of their users.
So, what do you think?
Do you think that the personalisation of content in news articles will predominantly be helpful or harmful? Or did you perhaps already encounter a situation in which the content of a static looking article or news page was changed, because of what this company knew about you as a consumer?
Because the implications of such a case can be huge, I am also interested in what you think that can be situations in which the personalisation of the actual content of news articles can be. How will this for example change the debates in the upcoming US Presidential election? Will news providers be able to change the actual news content, based on their readers’ voting preferences? And how will this influence the debate on consumer privacy and the legislation on personalisation?
I am looking forward to hear your thoughts!
Kind regards,
Jan Clerkx
Sources:
Bodó, B., Bastian, M.B., and Helberg, N. (2019) Personalised News – Implications for the democratic role of the digital media, user rights and public information policy. University of Amsterdam.
Chrysanthou, A., Barlas, P., Kyriakou, K., Kleanthous, S., & Otterbacher, J. (2020, March). Bursting the Bubble: Tool for Awareness and Research about Overpersonalization in Information Access Systems. In Proceedings of the 25th International Conference on Intelligent User Interfaces Companion (pp. 112-113).
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of general psychology, 2(2), 175-220.
Nielsen Norman Group (2019) The Dangers of Overpersonalization [Online], Assessed on 29-09-2020, via: https://www.nngroup.com/articles/overpersonalization/
The Verge (2019) Black Mirror: Bandersnatch Could Become Netflix’s Secret Marketing Weapon [Online]. Assessed on 29-09-2020, via: https://www.theverge.com/2019/1/2/18165182/black-mirror-bandersnatch-netflix-interactive-strategy-marketing
Thurman, N., & Schifferes, S. (2012). The future of personalization at news websites: Lessons from a longitudinal study. Journalism Studies, 13(5-6), 775-790.
Hi Jan,
Thank you for the interesting read! I think that you are completely right: over-personalization of news articles is most definitely a dangerous development. However, to some extent, newspapers have always been known to have a certain outlook on specific topics (e.g. Fox News has always been conservative, or – a Dutch example – de Telegraaf has always been sensationalist). Newspapers all target different segments, and have been doing so from the beginning, framing it in their own way. So, to some extent, opinions from people already differ depending on their preferred news source. I think that the principle of feeding oneself the information that fits best with your own view has existed for a long time, but now social media exists and newspapers are online, it has become increasingly worse.
I think that you are right to say that personalizing content further than that can be harmful, and it will require legislation, but that also will prove to be an interesting challenge, because free speech is also very important, especially when it pertains to the press. Therefore, I think it would be best to focus on what kind of data should actually be allowed to be collected. Voting preferences, in my opinion, should not be among them.
Hoi Jan,
Very interesting blog post! This is a phenomenon that I have noticed a lot lately, especially on less ‘formal’ news sources such as instagram and facebook. It has occured to me that when I look into a certain angle of the news, both news websites and social media keep on feeding me the news from that angle only. This frustrates me because I prefer to see things from all angles. Therefore I agree with you on your argument that personalizing newsarticles is harmful. It is important for people to read about things in a different lighting than they would usually do, so they can change their opinion if needed, which is not an unnecessary luxury these days.