-- MathewKenneally - 18 Nov 2014

FACEBOOK AND OK-CUPIDS CREEPY EXPERIMENTS: THE RULE, NOT THE EXCEPTION.

Revelations that social networking sites Facebook and OK Cupid conduct unauthorized and secret experiments has startled users. The response has focused on the immoral nature of the specific experiments. The experiments though are poignant examples of a systemic problem: that social networking has become monopolized by a small number of private businesses whose incentives are not to enable a more networked society but to generate profit by spying on and manipulating users. It demonstrates the necessity for alternatives to the current model of social networking.

In July 2014 it was revealed Facebook had altered the newsfeed of two groups of users. One group was shown positive status updates, the other negative status updates. The result was very slight changes: users who saw negative updates posted more negative material themselves and vice versa. Users were understandably indignant that Facebook had tried to make people sad.

Around the same time it was revealed OK Cupid, an online dating site, altered its algorithm to determine “matches” to describe what they considered bad matches to be good matches to test if users would accept the algorithm’s advice or rely on their own judgment.

Both companies raised similar defenses. Facebook apologized, but also insisted they were not trying to make anyone sad. Rather they were trying to test the theory that exposure to positive content on newsfeeds makes people depressed, and use the results to improve users’ Facebook experience. OK Cupid was more forthright, arguing without apology they were simply trying to learn how users responded to the site to improve its functionality.

Each defense falls back on the argument that websites were merely engaging in A/B testing. This is where a site owner direct groups of users to alternative versions of the website - A and B - to determine which works more effectively. Social networking sites, like online retailers, use this method as a matter of course.

This defense ignores key distinctions between social networking sites and other websites. Social networking is a form of social infrastructure. For many people, especially the young, participation in Facebook is perceived as necessary to maintain an active social life. Unlike a retail site, opting out is not an option. While there are more competitors, online dating is similar. Participation in online dating is becoming more widespread and increasingly the means by which people meet. Also, in certain age groups and markets individual dating companies may establish a Facebook monopoly. In either instance the choice not to use social networking sites requires a person to exclude themselves from aspects of social life.

Social networking sites, unlike online retailers or newspapers, have the power to shape our social reality. They can prompt us to contact one person over another, promote one event above another, and (potentially) impact our mood. OK Cupid is using individuals dating lives, unbeknownst to those people, to better understand the science of selling love. Further, innovations that integrate technology with visual stimuli, such as google glass, may enable social networking sites to alter what users physically see. Such glasses could display “Facebook profiles” or “dating profiles” of selected individuals within a user’s vision. The manipulation enabled by social networking is far more invasive and than simple A/B testing by a retailer.

The social networking business model is spying and manipulation. Their main source of revenue is users’ personal data that can be monetized in two ways: selling it to third parties; or targeting advertisements. Facebook has an incentive to experiment on users to increase the number of users, frequency of use, and effectiveness of advertisements. Some online dating businesses charge membership fees for use instead of selling data or advertising. Others sell user data to third parties and others rely on advertisements to support the site. The incentive for each is still to collect data and use it to advertise or increase its user base.

Social networking sites have an imperative to keep these processes obscure. For the data to have integrity, the experiment needs to be controlled. The users must not “know they are in the laboratory”. Users aware of data collection or manipulation might change their use or cease using sites all together. It is no accident these companies keep their code secret and bury privacy guidelines in incomprehensible terms and conditions.

There are proposed alternatives to the current model of social networking. First, decentralized social networking. This is where user profiles are not concentrated in a single server, but rather on servers run by individual users or a number of private companies holding communal servers. No one company or individual holds all the data or the capacity to alter users’ experiences.

There are some technological difficulties to be overcome in this model. The average internet user does not have an Internet server. If private sector bulk servers were to host most of the online profiles, data could again be concentrated in a few companies. Second, Facebook and online dating offers users a valuable feature: only Facebook, not other users, knows when you have viewed a person’s profile. Under a decentralized model each person hosting his or her own profile could track who visits it. We would no longer be anonymous to each other.

Another alternative is supporting ethical and transparent social networking companies. The platform Ello, like Facebook, operates under a client/single server model. However, Ello undertakes not to sell data, not to advertise, and only use data to improve the site. This model is imperfect because Ello, if a monopoly, would have the same power as Facebook. Further, Ello is not open source. This means the users cannot read the code themselves, which is essential to ensure Ello adheres to its own promises.

In the short term it is necessary to make users aware of the dangers of our current approach to social networking. The public’s disgust at Facebook and OK Cupid’s aberrant behavior can be used to generate interest and support for alternative approaches.