// Published 5th of February 2017
In western democracies we often pride ourselves for living in a society where free speech is guaranteed and all voices can be heard without interference by the government.
But with the rising popularity of social media sites like Facebook and Twitter there's been sort of a misunderstanding about what constitutes free speech. The understanding appears to be that the same guarantees which we have against governmental interference in free speech would just as well apply on social media sites. Which they don't.
As a matter of fact, the prevalent social media sites we have today are actually some of the last places where you should expect freedom from censorship or selective filtering. This is not a problem in itself - the webpages of Facebook and Twitter are the private property of their respective companies. You get a service for free which allows you to express your thoughts and ideas - however you can only do so at their discretion. The public needs to be educated on this relationship between a social media user and the respective company - otherwise outrage and confusion about selective filtering will make heated debates even more toxic.
Social media sites face a difficult trade off because they have to filter or delete content from their sites for various legal reasons: For example due to copyright infringements, threats of violence or promotion of Nazi propaganda (in some countries). In the past those reasons for deleting content appeared to be sufficient but recent developments on social media sites have prompted their owners to further scrutinize content. One well known example is bullying, which has become more and more common due to the perceived anonymity on social networks and the distant relationship between users (and the dehumanizing interaction on social media sites ). Twitter noted this new phenomenon of bullying as a reason for their new policy in 2015. During the 2016 presidential elections in the US another new phenomenon arose from social media sites: Fake news. Facebook recently announced a policy change which addresses the problem of virally spreading disinformation.
Both of these measures by Twitter and Facebook are perfectly legitimate for most of their intended use cases. However now both of these companies have taken upon them the role of an editor and users can no longer expect their social interactions (sharing of news stories, etc) to happen without any interference. While this might be unproblematic for most cases, there will always be corner cases (for example during a quickly developing news story) where this editorial interference by social media sites has unintended consequences - especially when the filtering happens in an automated fashion. Legitimate news articles will be removed (because at one point in time they appeared to be fake news) or false accusations of bullying cause important voices to be silenced. For example there could be a barrage of hateful tweets against an unpopular action taken by a member of the government - can an algorithm effectively distinguish between the bullying of a normal citizen and a government official?
Users of social media websites need to be made aware of these caveats. Social interactions on social media sites can neither be expected to be free of any censorship nor to accurately, impartially and equally reflect the opinions of all participants. A first solution to this problem could be a disclaimer which pops up as soon as a news feed of a user has been altered (for example when content from a person you follow is omitted in your timeline). Google already does something similar for searches if certain results had to be removed (often due to DMCA takedowns).
// 4th of May 2020: A discussion in the wake of the covid-19 pandemic: https://www.bbc.com/news/technology-52501453
This isn't a critique of social media sites. Due to the current state of interfaces and the way users interact with each other (mostly through text) these interactions are by design less "human" (for lack of a better word) than face to face interactions. This lowers the bar for personal attacks. ↩︎