Article by Dr Adrian Venables: Anti-social media – the rise in online censorship
Social media has been one of the defining technologies of the 21st century. Previously, Internet users had primarily been consumers of information, but these websites and applications enabled everybody to become content producers. The introduction of the iPhone in 2007 contributed to its growth by providing mobile Internet access and freeing users from the constraints of desktop and laptop computers. Combined with the launch of software optimised for mobile devices, numbers of users have steadily increased and is expected to exceed 3 billion by 2021.
The first decade of social media was a halcyon period in which the medium was regarded as harmless entertainment and was mostly free from state interference. However, as its power to inform, influence and alter behaviour became increasingly apparent, governments began to take a closer interest in the online behaviour of their populations. Although its exact contribution is still debated, Social Media played a role in the ‘Arab Spring’ of 2011. Popular with the younger generation, it enabled protesters to organise gatherings and spread news of events to a wider global audience. Initially caught unawares, affected governments responded swiftly by blocking access to social media sites and even temporarily severing Internet connections.
Whereas the result of mobilising populations in the Arab Spring were immediate, obvious and sometimes violent, another subtler effect was seen later in the decade. The 2016 Trump election victory is now infamous for the widespread and systematic online information campaign directed at the US population. This included what became known as fake news - outrageous news headlines designed to tempt readers to click on the stories and their embedded advertisements to generate revenue for their creators. More insidious though was the Russian government interference in the election, described by the 2019 Muller Report as ‘sweeping and systematic’. This was conducted through the St Petersburg based Internet Research Agency, which combined a range of techniques to generate a sentiment favourable to the Trump campaign. In addition to conventional paid advertisements, the Russians generated fake accounts purporting to be from US citizens. These online personas, termed sockpuppets, were used to comment on, promote or defend an issue. This was achieved by posing as a leader of a reputable group, reliable news source or trusted individual, which simulated grassroots support for Trump – a process termed astroturfing. These were supplemented with trolls; accounts set up to create disruption and division by posting provocative, misleading or pointless comments. In addition to the human operators, automated programmes termed spambots were also used to open accounts and generate traffic to develop online engagement. By clever use of hashtags, it was possible to manipulate the algorithms used by social media to dominate online discussion.
Although social media is a technology based medium, it harnesses some very human characteristics, which can be employed to manipulate and influence users. These were utilised in the 2016 US election and are now being widely employed to censor and control behaviour. The first of these is homophily, which is the tendency for people to have ties with those of similar beliefs. Social media users tend to self-censor by only associating with those of similar views and accessing news outlets that promote stories that do not challenge existing viewpoints. In time, this leads users to only be exposed to opinions that coincide with their own. This reinforces their opinions as being the ‘right’ or those of the majority and does not allow alternative ideas to be considered. This comfort zone is termed an echo chamber and can result in a very narrow perspective of an issue with social media and some search engines contributing to the process. First termed as the filter bubble by Internet activist Eli Pariser, this is the situation that occurs when website algorithms selectively provide information based on past browsing behaviour. This can be illustrated by comparing the returns from similar inputs to different Internet search engines and how Facebook’s personalised news stream differs from other news sources. As beliefs become more entrenched, confirmation bias can emerge. This is the situation in which individuals favour information that aligns with their preconceived knowledge, even if flawed, and choose to disregard alternative opinions.
Homophily, echo chambers and confirmation bias are human traits, and yet even for those seeking alternative perspectives, online censorship and manipulating is increasingly preventing access to some opinions. Recent research by Northwestern University in the US highlighted a potential bias in Google’s search algorithm that favoured predominantly left leaning news organisations in their rankings. This is particularly significant as online news sources are gaining prominence over traditional media organisations. Moreover, social media now often pushes news items to users, who may not actively seek other information sources. Combined with the prevalence, ease of access and convenience of online resource, consumers may be subject to unconscious bias and censorship without their knowledge.
For those who do wish to access alternative views and form their own opinions, the range of online resources available may be increasingly limited. Following the role that fake news played in the 2016 elections, governments and news organisations are increasingly citing it as the reason for censoring and removing material. Whereas attempts to verify and confirm the factual content of stories are admirable, there is a danger that their definition of fake news will spread to unpopular or divisive news. Some countries such as China are well known for its authoritarian control over the Internet within its borders, but others are also seeking to control what may be posted. Russia has recently introduced a new law that could effectively disconnect its Internet from the rest of the world. Justified as ensuring resilience in case of cyberattack from abroad, by directing all traffic through centrally controlled routers monitoring and filtering of information originating outside Russia could also be implemented. The EU’s new copyright laws, which applies to social media companies have also raised concerns. With 19 nations voting in favour with 9 including Estonia voting against or abstaining, the law is intended to bring existing regulations into the online age by making Internet platforms liable for content uploaded to their sites. Licenses must be obtained from rights holders for copyrighted works to be hosted with filtering used to remove unauthorised material. Critics have stated that this is impracticable and unworkable and will result in online expression and free speech being curtailed. Faced with prosecution some believe that internet companies will take the safe course of action and will remove the majority of images and media currently available online.
Freedom of expression and the issue of online free speech is becoming closely related to censorship and is an increasingly contentious issue in western democracies. Opinions and views vary but with the Internet’s infrastructure and websites owned by either governments or private companies, consumers have little influence over how it operates. In May 2019, Facebook removed a number of prominent conservative figures from its platform labelling them as ‘dangerous’. Critics were quick point out that several far-left activist groups openly advocating violence remained active. European governments are also active in policing online content. The UK is very active in this area with specialist units devoted to monitoring social media. Using the justification of investigating ‘hate crimes’ social media activity can be sufficient to attract police interest if a post causes someone to be offended on a range of issues. This has effectively muted many forms of debate and criticism on a range of contentious issues including gender and religion. Restricting freedom of expression and free speech may lead to what is termed a spiral of silence. Proposed by political scientist Elisabeth Noelle-Neumann in 1974, the term relates to the tendency of people to remain silent on an issue when they feel that their views are in opposition to the majority. Today, an individual may remain silent and feeling prevented from expressing an opinion online for fear of being accused of a ‘hate crime’. In doing so it deters others from stating a similar view and leads to the views of a silent majority being supressed by a vocal minority.
At the opening of the Estonian Riigikogu (Parliament) on 25 April 2019, President Kersti Kaljulaid wore a sweatshirt with the slogan 'Sõna on vaba' (the word is free). This commitment to free speech was again emphasised during a meeting with the European Federation of Journalists a month later. Freedom of speech is binary – you either have it, or you do not. With complete freedom of expression is the acceptance that those you disagree with, including extremists of all persuasions, will have a free platform. Once limitations are imposed, the challenge is where to draw the line and accept the risk that the restrictions may increase over time. It will be for future generations to debate whether the free speech permitting, pre-2016 Internet and its social media applications was better than what it subsequently became. That is of course, if they will be allowed to debate such issues online.
Author: Dr Adrian Venables, TalTech Centre for Digital Forensics and Cyber Security senior researcher
The article was published in Edasi.org.