Taming the Beast

Guest post by Meghna Yadav

Every April Fools Day, the administrators at Reddit come up with a social experiment, a game designed to be played by the members of their 9 million strong community. 2017 saw ‘The Place’, a (999,999 x 999,999) pixel board, where you were given a 16-color palette and could put a single pixel down every 10 minutes. Working on a project individually was almost impossible. The Place had all the hallmarks of a classic collective action problem and in fact, Reddit administrators put it best- “Individually, you may create something. Together, you may create something more”. The game was played by more than a million online strangers over 3 days. And even though there was chaos in the beginning with everyone trying to make individual pieces, people soon banded together to result in this-

The staggering amount of art that was achieved (including pixel versions of Mona Lisa and The Starry Night) can be seen once one starts zooming in. Surprising also is what did not appear. Your regularly scheduled trolling, abuse, and offensive iconography never appeared for a significant amount of time, while political activism too never got much space. What jumps out is that the most popular interests on the board were national flags (close to 30% of the area). In a community such as Reddit, defined more by their anglophonic libertarian ideology, nations continue to have wide appeal. Schelling would ask us to please not assume that this means all the thousands who worked on bringing these flags to fruition wanted to make the flags. Simple ideas such as national flags have wide appeal and are easy to implement. So is that an indication of how people coordinate over the internet? Is the internet susceptible to generating action for simple widely appreciated ideas? These are the questions we address in this paper today.

From way back in the 1990s, when the Zapatistas- a largely Mayan group in southern Mexico- advocated their cause in an innovative manner- through email lists and usenet groups, to the recent Ramjas protests in Delhi University that saw a nation erupt over Twitter and Facebook, that the internet has significantly helped collective action cannot be denied. However, the kind of arguments that rely on a technology-deterministic perspective to label these movements as a product of the Internet- “Tweets were sent. Dictators toppled. Internet = Democracy, quot erat demonstrandum” aren’t correct (Morozov, 2011). During the Arab Spring of 2011, the protests were quickly labelled a Facebook or Twitter revolution. Such a perspective, propagated often by Western media overemphasizes the role of the internet and can be dangerous. So, while social media is an important platform for aggregation of information and coordination for action, some strong questions about its structure and inherent biases need to be asked. The study of internet and collective action is multidisciplinary at best but we’re going to try and see what Economics has to say about it that goes beyond debates of e-commerce and net neutrality. The way I see it, there are two interesting questions that can be addressed in this paper- a) On what lines are people aggregating over the internet and why, b) what kinds of aggregation translate to real change.

Selective Exposure and Segregation over the Internet?

Ever log on to any of your social media profiles and find everyone talking about the same thing? Social media seems to favor certain ‘darling’ issues which change on an almost daily basis. This is worrying because while there is a lot of information floating around on the internet, what most people run into is often the issue du jour. An RCT on how social information environments affect collective behavior by Helen Margetts et al (Oxford Internet Institute) showed that participants were significantly more likely to sign a petition if it had a high number of other signatories (as opposed to medium or low). It also showed, quite interestingly, that certain petitions faired better than others (for eg, Fair Trade faired the best and Rights in Tibet the worst) bringing attention to how the average internet user is fed a certain narrative and particular interests are better represented than others.

Does this have to do with the design of social media platforms or the design of its users, us common people? The answer is a combination of both. Any model of information cascades gives us a good idea of how certain stories (including ‘fake news’) can spread like wildfire over social networks of family and friends. When we take this model online where social media algorithms are designed to provide a very real structure to ‘echo chambers’ (what you ‘like’ determines the content you are shown which is often material related to your ‘likes’) thus compounding the effect of these information cascades. Social media is also structurally weak to the threat of bots and trolls. As Philip N Howard of the Oxford Internet Institute calls it, this ‘Computational Propaganda’ is designed to muddy the facts, with the aim of making it difficult for neutral bystanders to discern the truth and easier for partisans to reject any views that may clash.

The same goes for self-selection and segregation models as propounded by Schelling. The internet is particularly susceptible to selective exposure and confirmation bias. While the Internet is technology that can theoretically let two very diverse individuals from opposite parts of the world have a conversation, it is hardly the case that it is known for its avenues for meaningful conversation. The way that the internet is monetised currently, companies like Facebook, want people to spend lots of time on their websites and to share/like lots of pages. For this, they need to create a safe space for the users leading to shielding. Where the internet offers anonymity, conversations often devolve to crass abusing and shouting matches and where it offers visibility (on social media), the tension between better information aggregation and reputational motives often leads to the creation of echo chambers. Scharfstein and Stein’s ‘Reputational Herding’ model applies very well in this context. By posting your true opinion and signal on social media, you can better the informational content but by posting the more widely appealing opinion, you gain in reputation. And so like in our fishery with two lakes, every time you echo the popular opinion, you gain in likes (higher average product) but the social network loses out on information. Thus, the lake with the popular opinion is overfished.

All of these problems exist even in the absence of the internet, but it is the design of social media that has led to higher gains in reputation with large social circles and lower costs of making opinions known, leading to a compounding of the problem. Selective exposure is alright when I am talking about how I enjoyed the latest Star Wars movie on a forum, but not when you are critically discussing demonetisation.

The wide reach of the internet has led to an illusion of being able to connect with diverse audiences and it’s good at feeding this illusion. This is especially true for developing countries where the internet penetration isn’t very high. When nationality continues to be one of the attributes people identify strongly with, talking with someone from say, Sweden gives the illusion of a differentiated audience. We forget about the digital divide and the way social algorithms function. The chances are high that the reason you are able to connect with someone from Sweden is because you are similar in many aspects, as many urban anglophonic youth nowadays are. The truly differentiated audience you seek might just be in an underdeveloped area of your very city but those would be people you would never normally talk to.

Does Internet Activism Translate to Real Change?

The ‘Slactivism’ hypothesis would say no. It contests that convenient and easy-to-access online services lead to a jump in slacker activism ie low-cost and low-risk online collective actions such as signing online petitions on change.org or “liking” a group’s Facebook page. One could look at this through the Benabou-Tirole ‘Incentives and Pro-Social Behavior’ and ‘Motivation’ models, but with now easy access to high reputation at lower costs which would imply lower sustenance of collective actions; when users satisfy their need to participate through clicks and shares, they would be less willing to join subsequent protests. In her book on ‘How the Internet Shapes Collective Action’, Sandy Schumann gives the example of the KONY 2012 campaign, where despite 3.7 million people pledging support, barely anyone participated when posters had to be put up all over the globe. However, in a recent study of a candidate winning the Mexican elections for state governor solely on the basis of social media campaigning, researchers try to refute this hypothesis. What is essential to note here is the difference in the two types of civic engagements in these examples, the main one being in the degree to which people’s own lives (their own utility functions) were affected by the collective action. So when the action required relates to the makeup of our governing civic bodies or is to assert our right to free speech, the internet seemingly works wonders. But when it is to do with creating awareness about children dying in some country that doesn’t figure in our day-to-day lives, the Slactivism hypothesis may turn out to be true.

In a recent article in The Hindu, Keerthik Sasidharan talks about the need for a citizens’ movement to make the Swachh Bharat Abhiyan a success. He is critical of social media campaigns as a one-stop measure to galvanise the public. He points out that Mr Modi’s token online posts of celebrities cleaning the streets are public theatre and will not be transformed into something more meaningful unless the symbolism is sustained. The UNICEF Sweden Director of Communications, Petra Hallebrant, expressed this more clearly when she said “Likes don’t save children’s lives” (Khazan, 2013).

The internet itself has been transforming- from the static Web 1.0 to the more dynamic, interactive and user-content oriented Web 2.0. As we translate more and more of our lives to the virtual world, the design and regulation of this world needs to be questioned. A key examination we must carry out is of the nature of collective action that is successful versus those which are not, and how, if it is possible, to make them successful. Since our analysis has been of the interaction of social models with the structure of the internet, the answer once again, must lie in re-designing the internet to effectively fulfill its role as an information broker and providing people with nudges delivered through social media to incentivise required behavior. We need to ask whether internet giants like Facebook and Twitter should be given a pass on the normative obligations for democratic discourse that we hold other media to.  Social media is increasingly becoming an important secondary, if not primary, source of information for many; a discussion on the creation of a better infrastructure of the same is necessary.

(Meghna Yadav is an MA student at the Delhi School of Economics (2015 Batch))

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s