Googlelections How Internet gatekeepers can manipulate the democratic process
- Political communication is immensely important and subject to bursts of innovation;
- There have been a number of waves of innovation, most recently with President Obama’s use of social media;
- The utopian dream of a totally free Internet is not a reality for the majority of its users, who are prisoners of comfort, habit and coordinated narratives;
- Internet media companies have begun to act like their traditional media counterparts – as gatekeepers for information, with the power to make or break elections, by manipulating perceptions, access to information and emotions;
- They have especially been adept at alternately facilitating and hindering the spread of user created political content, which has made the Internet into a new public commons for political debate;
- Examples given include Google, Facebook, Twitter, as well as academic interpretations of some extraordinary involuntary experiments run on users;
- Everything is in transition, including the current Internet giants, who face disruptive innovation from nimble challengers and issues with their business models and profitability;
- A new model of political communication is also emerging, centered on communities of prosumers, consumers and producers of energy, who evangelize for certain issues and candidates as part of grassroots reactions against an establishment perceived as biased and unfair.
Innovation in political communication
Throughout history, the relationship between rulers and those who ruled was, by necessity, a distant one that, nonetheless, had the effect of enabling propaganda of divine origin or the act of being anointed by God and the imagination of the people to create larger than life figures out of mere men. Then, as with so many things in our age of modernity and post-modernity, America changed the paradigm. President Franklin Delano Roosevelt became engaged in a fundamental reworking of America’s relation to government and business in the New Deal era after the Great Depression and then in mobilizing the American people for entry into WW2. He chose the medium of radio to speak directly to the citizens and, in so doing, also spoke to the electorate, and thereby changed the rapport between the American President and the people. He did not abuse the new power, because he was a skeptic with regards to the effectiveness of too much personal leadership as opposed to infrequent, but memorable, addresses. He made only 30 of his “fireside chats” between 1933 and 1944, but captured the American imagination. Soon enough, many political leaders started using radio as the principal means of effective mass communication. King George the VIth of Great Britain gave the first radio address of the monarchy in 1939, explaining the declaration of war on Germany, as presented in the 2010 movie “The King’s Speech”. But it was not just a tool for democracies, but for autocrats as well, who need mass communication to strengthen their personal positions and to reduce the apparent distance between the people and power. This lessens the impact of the absence of representative structures which endow their rule with legitimacy by creating the illusion of proximity.
In Joseph P. Berry’s book, “John F. Kennedy and the Media: The First Television President”, he recounts how JFK leveraged his good looks and demeanor to project an image that would capture American people at the dawn of the age of mass TV ownership. He trounced his opponent, Richard Nixon, in the first televised presidential debates not necessarily on the weight of his arguments, but on the strength of rhetoric and the physical projection of the attributes of idealized leadership that presaged America’s later infatuation with its “Camelot era”. JFK became, over the wishes of his advisors who felt that mistakes or over familiarity with the press would undermine the Presidential Office’s dignity (which they did), the first President to give unscripted and free flowing press conferences aired directly to the American people. For the entirety of his short presidency, he held a press conference, on average, every 16 days, with the aid of the newly established White House Press Corps. The first had been viewed by 65 million people and a poll taken in 1961 showed that 90% of those interviewed had seen at least one of his press conferences.
The fact of the matter is that the time when President Kennedy started televised press conferences there were only three or four newspapers in the entire United States that carried a full transcript of a presidential press conference. Therefore, what people read was a distillation. . . . We thought that they should have the opportunity to see it in full. Pierre Salinger, Press Secretary to President Kennedy, John F. Kennedy Library Oral History Interview
The next wave of innovation truly came with the ascent of President Barack Obama [1], who chose an electoral communication and fund-raising strategy that supplemented traditional media with the Internet and social media. He did this to attract the younger generation of voters, who were traditionally less interested in elections (and traditional media), and also to capitalize on his message of hope and change. The consumption of Obama as the herald of a Camelot for the diverse era relied on Barack Obama’s exotic origins and deafening but unspoken appeal to a rejection of America’s history of tense race relations, a reductionism into which every American is, whether he wants it or not, indoctrinated.
Barack Obama’s 2008 online performance (Source: Edelman) (Note – his Facebook follower list grew to 20 million people for the 2012 campaign)
The use of the Internet promoted the same psychological effect of familiarity or a closer rapport that prior innovations had, but to which the American people had become largely inured over time. The television stations were largely seen as a new intermediary between the people and the leader, rather than the factor of disintermediation that their parents had considered them to be. Familiarity breeds contempt in a world of unending pursuit of novelty.
The novelty of new methods of communication attracted trendsetters and gave them a freshness that, in the “Wild, Wild West” of the Internet, seemed to be truly free of third party manipulation. However, this was just an appearance, and the recent electoral campaign in the United States has proven that innovation is no match for corporate consolidation. Just as in print media or traditional broadcast media, both business and community consolidation took place under the leading digital companies. Networking effects meant that not just disaggregation into granular communities tied together by the Internet became possible, but also mega-networks. A virtuous cycle of high number of users raising the lure of a certain network guaranteed that they would eventually come to dominate the Internet media landscape, without even the veneer of different brands and publications that hegemonic media companies maintain as a façade. Barack Obama was everywhere on the Internet, but he was present through only a key few services that provided the greatest coverage.
These, in turn, like William Randolph Hearst’s newspapers, the major radio broadcasters, the major TV stations and the major media companies before them, became gatekeepers for the new generation of political thought and speech.
The gatekeepers
A gatekeeper is somebody who controls access to something and yields extraordinary power because of this, which he can use to help his friends and hurt his enemies. Left unchecked or even unsuspected, gatekeepers can mastermind significant abuses. A gatekeeper may decide who gets into a certain school, who gets a job and who gets his views heard. This eventually leads to political power. American political scientist Paul Gottfried’s book [2], “The Great Purge: The Deformation of the Conservative Movement” describes how William F Buckley’s control of the National Review, the flagship publication of the American conservative movement after the 1950s, gave him the extraordinary power of defining who was conservative and who was not. The frequent “weeding of his garden” was done not just to eliminate cranks and dangerous characters in favor of worthier thinkers (a role of gatekeepers), but also to purge currents of thought that he did not agree with, or he did not agree with anymore or conflicted with the direction he wanted to set. His prestige among elites and the rank and file of conservatives led to a bottleneck where those he purged were shut off completely from the movement and became marginal figures, rather than competitors in their own right. Among them was the John Birch Society, a stridently anti-Communist organization that alleged a Communist infiltration in all areas of government (they were proven right in principle, though maybe not in degree). Others were more nuanced. He eliminated, for instance, the Israel-sceptic conservatives who feared the loss of American prestige and influence in the Muslim world, where it had been seen as an honest broker in the first part of the XXth century, as described by Eugene Rogan in “The Arabs”.
Control over multiple gatekeepers or even an entire hierarchy of them leads to the accumulation of power and influence. In autocratic regimes, almost all gatekeepers eventually answer to the autocrat, while still exercising power in their domains. In free countries and democratic regimes, the gatekeepers may be outside the bounds of centralized control, but that does not mean that they are any less dangerous and the danger comes not from some sort of illegitimacy or moral turpitude of their role, but through their very legitimate and protected status, including through free speech codes. Newspaper and television station owners may decide what views they promote; private schools (more so than public schools) decide who they let in and what the criteria for selection will be; private venues decide what events they will host etc. This is a part of pluralism, with ownership acting as a shield against abuses and a homogenizing central control. In theory…before the great aggregation of media companies. In the United States, 90% of media outlets are controlled by 6 media companies. The remorseless logic of the marketplace was at work, but the efficiencies gained through aggregation also led to the emergence of the ability to impose narratives that shape reality for those immersed into them, aided by a warped grasp of basic facts and a reflexive belief in journalistic objectivity.
Every bottleneck must eventually be broken by dissidents, which, in capitalism, meant the appearance of new technologies and mediums for communication. The Internet was and still is such a place, but most people never actually experience the Internet and its treasure trove of knowledge and dissent. Like Apple users discovering the limits of their market for apps, the Internet that most people experienced is a “walled garden” curated more or less lightly by the owners. They can choose to enable maximum freedom, they can choose to restrict, or they can choose to intervene from time to time, sometimes in an arbitrary fashion. Apples, for instance, after the American moral panic regarding the Confederate battle flag and the subsequent push to eliminate it from the public sphere (by dint of guilt through association with the Dylan Roof killings), eliminated numerous instances of the battle flag from the app store and even its users’ devices, including strategy games set in the Civil War. Another inkling of the latent power of such companies came when a licensing dispute for George Orwell’s novel 1984 led Amazon to not only retract it from sale, but also delete it from its users’ Kindle reading devices.
Moral hazard
All is not well in the anarcho-libertarian Internet we were promised. Most people cannot even imagine the extent of the intermediation of their Internet experiences. The main medium for searching for information and products online is Google search, so much so that it is becoming a generic term like Xerox. The Computational Media Lab at the University of Maryland states that “search engines are powerful brokers of political information that can shape an electorate’s perceptions. Political biases can emerge from the complex and opaque search algorithm, surfacing more or less supportive or critical information about candidates”. Facebook has over a billion users, who increasingly use it as their main source of news and opinion. Twitter has redefined (for the worse, some would say) brief and catchy communication. Less relevant, but just to illustrate the power of market leaders in the networked age, the top three browsers, our main software for accessing the Internet, command a 95% market share.
These examples do not specifically produce content, but they act as our main intermediaries for consuming the content produced by others, whether other users or other media companies. That their power is enormous was accepted fact, and proven by the 2008 elections. That they would be active gatekeepers, exercising their power on behalf of one ideology or another or one politician or another, was less apparent in 2008, because everybody, including Silicon Valley, was on board with the Anointed One, Barack Obama.
The reality became apparent in 2015, with the preamble to the American elections, and reached its apogee during the bloody political fighting of summer and autumn 2016. The tech world became an active combatant against the candidate preferred by around half the electorate, Donald Trump, with only entrepreneur Peter Thiel (who does not control an Internet media voice) dissenting in favor the Republican candidate. In the hysteria surrounding the acrimonious debates, with charges of Nazism, racism and sexism bandied about to the exhaustion (or riot) of the electorate, Silicon Valley started asking itself whether or not, as geeks would say, “with great power comes great responsibility” to prevent the disasters that the “Trump as Antichrist” rhetoric predicted. This was an actual topic of discussion during internal meetings at Facebook (it was in the top 5 request for clarifications), for example, as leaked to Gizmodo.
The intermediated dystopia
The influential Politico Magazine in the United States ran an article by psychologist Richard Epstein in August 2015 [3], weeks after Donald Trump announced his candidacy to a storm of outrage and derision, titled “How Google could rig the 2016 elections” and subtitled “Google has the ability to drive millions of votes to a candidate with no one the wiser”. This is in a country where the 2000 elections were decided by a few hundred voters in Florida and the 2016 elections were ultimately decided by slim majorities of thousands of votes in battleground states. He quotes a study he conducted [4], “The search engine manipulation effect (SEME) and its possible impact on the outcomes of elections”, which showed that simple changes to Google’s search ranking algorithm can shift the voting preferences of 20 to 80% (for some population groups) of undecided voters (a bit over half the American electorate at any one time). The article’s abstract states:
“We present evidence from five experiments in two countries suggesting the power and robustness of the search engine manipulation effect (SEME). Specifically, we show that (i) biased search rankings can shift the voting preferences of undecided voters by 20% or more, (ii) the shift can be much higher in some demographic groups, and (iii) such rankings can be masked so that people show no awareness of the manipulation. Knowing the proportion of undecided voters in a population who have Internet access, along with the proportion of those voters who can be influenced using SEME, allows one to calculate the win margin below which SEME might be able to determine an election outcome.”
Moreover, given the polarized nature of many elections throughout the world (including Romania), which are decided by slim majorities (even leading one former Romanian Presidential candidate to publicly express his undying love to his wife, thinking that he had won, only to have victory snatched from under his nose), Google can actually decide 25% of world elections. In the United States, the margins for half of electoral victories have been under 7.6%, and the 2012 election was won by a margin of only 3.9%. An algorithm, especially for objective measurements like search history, can be very close to true objectivity, though there is significant dissent which will be covered in a future article, such as the book “Weapons of Math Destruction” by Cathy O’Neil (reviewed here [5]). Google, however, admits to adjusting its algorithms (the “secret sauce” of their success compared to Bing or prior search engines) at least 600 times a year. An article [6] by researchers from the University of Arizona published by Slate, a left leaning magazine, stated the following, through they argued that the political favoritism was an unplanned, emergent property:
“As early as 2010, researchers at Harvard University started finding evidence that Google’s search rankings were not so objective, favoring its own products over those of competitors. A Federal Trade Commission investigation into the conglomerate in 2012 also indicated evidence that the company was using its monopoly power to help its own businesses. So it’s no secret that Google search results aren’t a font of objective and unbiased information.[…] Google is not fair; it favors some candidates, and it opposes others. And so far, it seems to prefer Democrats.”
- That you can change the proportion of people who favored any candidate by between 37% to 63% after just one search session;
- Giving three groups of people 30 search results and the opportunity for a quick browse led to significant changes of support for one fictional candidate or another. The difference – each group had a different ordering of the same articles, which determined who you supported;
- In a real life experiment during the 2014 Lok Sabha elections in India, a 2,000 person group had its voting preferences shifted by 20-60% through selective ordering of information.
The authors support the idea that SEME voting preference shifts are at work in India, with the eventual winner of the last elections, Narendra Modi, appearing in searches 25% more often than his rivals in the months before the election, which could be easily influenced by a better position in search rankings for various issues. The chart below, taken down by Google, but saved by the authors beforehand, shows this.
The illusion of neutrality helps, as well as the fact that more and more people are going online to find out about the issues. A 2012 Pew study [7] showed that “73 percent of search engine users say that most or all the information they find as they use search engines is accurate and trustworthy”. Search engines are more trusted than the news media itself, which is why 50% of clicks go to the first two results, with more than 90% of all clicks going to the first search page.
Another area of possible manipulation is the autocomplete options when inputting search terms, as well as the snippets of information Google collates on the right hand side of the screen when searching for information. The picture below is from another article [8] in Slate written by University of Arizona researchers addressing this issue. The article also quotes the Computational Journalism Lab at the University of Maryland regarding differences in the way candidates are portrayed – unequal space given to each candidate, the use of quotes that do not adequately explain their stances and overreliance on a limited set of news sources. The algorithms may as well be objective, as long as the entire media field evinces a particular bias.
I personally experimented with Google for a bit and came up with this result.
As a GM CEO, Charles Erwin Wilson, once said, “What is good for General Motors is good for America”. Google and other companies may also arrive at a similar conclusion. Google is especially suspect, given its power, its ideological and financial ties to liberal interests most often expressed through the Democratic Party and opportunistic Republicans, and its contributions to recent Administrations (eight of the Obama tech officials were from Google). Those same appointments also facilitated access to the White House for Google representatives, according to a report by the Wall Street Journal, with weekly visits leading to ten times more exposure than other comparably sized companies [9].
Other examples
Google, of course, is not the only watcher who must be watched. There is already a long list of academic and internal research showing the power of Facebook in promoting not just products, but issues and political candidates:
- In 2010, an experiment [10] in social mobilization on 61 million newsfeeds (the presence of an “I’m voting” button) led to 340,000 extra voters turning out for the US Congressional elections;
- During the 2012 elections [11], Facebook ran an experiment on 1.9 million newsfeeds to see if increasing the amount of hard news can improve voter turnout;
- Facebook also ran an experiment on 700,000 people, described academically here [12] to see if increased emotional content on newsfeeds can trigger higher sharing levels and “emotional contagion”, whereby lack of direct contact with friends is not a barrier to the transmission of emotional states.
Of course, if it wanted to, Facebook could ban all content related to a candidate or an issue or simply make it unable to travel through its network. It is a private company and only direct affiliation to a candidate can lead to restrictions on its activism. People might understand that traditional media can have biases, but they do not expect that the intermediary for user created media would purposefully curate the transmission of knowledge, irrespective of legality and end-user preference. There have been numerous instances of news failing to spread and trending stories of a certain bent (such as reports on the crimes of migrants in Germany) not appearing on the trending interface which facilitates viral dispersion, despite having the raw numbers to be placed there. The infamous communication captured on microphone between Chancellor Angela Merkel and Mark Zuckerberg of Facebook regarding the regulation of content critical to refugees and the country’s open policy towards them also shows a new level of activism on the part of tech giants who are not simply satisfied with wealth, but have the means and desire to pursue their political preferences, even if they are utopian.
Youtube, which belongs to Google, also makes it a point to utilize its relaxed copyright infringement policies, which place the burden of proof on the accused, in order to rapidly take down political material and limit its viral potential, even if the video is ultimately reinstated.
Wikipedia, along with the decrease in its base of contributors and editors, has witnessed the rising power of influential editors to curate political content on the website and silence dissent. For instance, the current nominee for Secretary of State in the future Trump Administration, Rex Tillerson, had his page modified immediately after his nomination to include many references and photos to his relationship with Vladimir Putin. Conversely, just before Hillary Clinton’s use during a debate of former Miss Venezuela and recently naturalized citizen of the US, Alicia Machado, to shame Donald Trump for his alleged behavior towards her, Alicia Machado’s Wikipedia page was heavily edited [13] to remove descriptions of her past scandals, such as infidelities, sex tapes, allegedly being the getaway driver for an assassination, allegedly bearing the baby (now a US citizen) of a noted Mexican drug lord of the Los Negros cartel and threatening a Venezuelan judge [14]. This kind of editing [15] can only be done through the connivance of highly place editors [16].
A very interesting form of censorship, which is practiced especially on Twitter, but also on Youtube and on Facebook, is “shadowbanning”. Rather than banning a particular user, you let him stay active and thinking that he is getting his message out, while affecting the number of his followers who receive his posted content. There have been instances of posted videos not showing up in the feed of Youtube subscribers, or of Facebook posts not appearing on the walls of users. Donald Trump, himself, as Republican nominee for President, allegedly experienced shadowbanning [17] on Twitter, with his tweets not going out to all of his followers.
Seeds of future disruption
Ultimately, a reckoning will be at hand which will limit the power of the current social media giants.
Firstly, they can become so useful to whatever ruling regime they serve that their own nominal independence is affected, as the past beneficiary of their influence seeks to cement control of them on account of his critical dependency. He is basically bringing gatekeepers under his direct control in order to manage their power. China proposed this to Google and Facebook, which refused at the time, even though their current practices are much closer to the exigencies of the Chinese social media control standards [18], which is why a number of American social media platforms are finally trying to enter the Chinese market. The key to bringing them to heel is dangling financial gain in front of them, either through access to a market of over a billion people (which matters, given the low and decreasing levels of monetization of individual users) or some other means of making money, such as permissive laws on the monetization of private data.
Secondly, the recognition of their influence may lead to their classification as public utilities and critical infrastructures, bringing them under the heel of national legislation dedicated to free and unfettered access to the public commons for the purposes of political speech. They may even become nationalized and regulated as “public broadcasters”. This is not as unlikely as it sounds in our neoliberal age. The weakness of many tech giants is that they find it difficult to monetize their network without alienating users and giving rise to their competitors (does anyone remember hi5 and myspace?). Therefore, their business model is woefully lacking and, consequently, income streams disappoint tremendously compared to the market valuations that exuberant IPOs assign to them. The situation is also unsustainable – there is a limit to how long the market will bear an overpriced, underperforming asset. The messaging app Whatsapp was sold for a significant sum rather than trying to become viable on its own. Twitter may face the same problem, After the Trump shadowbanning scandal, AOL gave up on buying Twitter, which hurt it pretty badly. Therefore, nationalization may become viable. The owners of overperforming stocks with underperforming fundamentals would likely be open to the fair market price reimbursement policies under eminent domain, especially if negotiation can sweeten the price.
The owners of overperforming stock with underperforming fundamentals would likely be open to the fair market price reimbursement policies under eminent domain, especially if negotiation can sweeten the price. |
Thirdly, there is the option of simply creating competitors to these entities to give a viable alternative that meets whatever standards the original companies did not meet. The founders often look forward to disrupting the entrenched tech giants with better features or more user-friendly policies. China created its own versions of services that did not come to the country or did not accept the terms and conditions for operating in the Chinese market. Those competitors, walled off with over a billion potential users, had the opportunity to grow and develop, becoming, in the future, competitors for Western companies. Sometimes, the applications are even better due to the rapidity with which innovations can be widely adopted on a top-down basis. This author, due to frequent working visits to China, uses Wechat, the Chinese version of Whatsapp. You can order off the menu in a restaurant with it, pay your bill, and other apps even have simultaneous machine translation for chats, which are very useful in certain situations. It has come to the point where Yahoo’s stake in the Chinese company Alibaba [19] (inspired by eBay) is worth more than Yahoo itself, whose mail hosting services have had numerous technical and security issues, meaning that Yahoo as a company has negative market value and is in fact worth less than the sum of its parts, or even just one of its parts [20].
On the Western “front”, a group of political activists who supported Trump and other dissidents have launched a “forking” effort [21] to create competitors to current social media giants and other Internet infrastructure. Mozilla founder Brendan Eich, who was removed from the Foundation’s board after it was revealed that he had donated money to anti-gay marriage political action groups, has created a new browser called Brave [22], now in Beta, which is geared towards protecting personal data and improving access speeds by blocking background programs from websites. Publisher and philosopher Theodore Beale (also known as Vox Day) was directly involved in the rapidly [23] growing [24] “alt-tech disconvergence movement”, with gab.ai [25] as an alternative to Twitter (400,000 people waiting list for an account) and a Wikipedia dynamic fork (meaning that they took a lot of its content, and many of its editors, but have different editing policies) called “Infogalactic – the planetary knowledge core” [26]. It remains to be seen whether they will succeed in becoming self-sustaining or influential. The high capital costs for entry to the field are why, for example, Youtube will be among the last major Internet infrastructures to be disrupted in this way.
Finally, it is possible that the current Internet media giants will fall prey to disruptive innovation and be added to the list of the dinosaur media that they used to deride. Their business models are certainly not improving visibly enough. Google, for all of its Alphabet innovations in driverless cars and anti-aging research, has yet to make money from them and history is rarely kind to first arrivals in a new market (Microsoft’s failed PalmPilot was the main touchscreen device before Apple took the market by storm). Google makes the bulk of its revenue from online advertising, which also fuels the Internet. Consumers have become inured over time to ads, and sometimes even actively hostile to the intrusive sort of ads, so the model itself is bringing less and less business to the ad buyers, the clients, putting pressure on the entire Internet establishment. Even Facebook, with its successful IPO, is getting desperate, launching new services and then inflating their results for media, markets and potential customers. A Bloomberg [27] article states the following:
“On Friday, Facebook revealed faulty metrics with Instant Articles, its mobile publishing system, the fourth disclosure of a measurement error since September. The admission sharpened calls for more independent organizations to monitor the performance of digital advertising. And some large firms that buy a lot of ads said they will more closely scrutinize their spending on the social networking giant and could shift marketing dollars elsewhere.... In September, Facebook shared its first measurement error: inflated viewership numbers for its video ads, a relatively new product. Two months later, the company disclosed additional metric errors along with new tools for third-party measurement companies, including ComScore and Nielsen, to track its system more closely. Problems persisted. Earlier this month, a report in Marketing Land, an industry publication, spotted a discrepancy between Facebook's internal metrics on how articles where shared and public measurements. Facebook confirmed the error. ‘That shouldn't happen,’ said Brian Wieser, senior analyst, Pivotal Research Group. ‘If anyone was concerned that Facebook's self-audit was not sufficient enough, they just proved it.’”
Already, we can see glimpses of what follows in terms of political communication. Donald Trump’s campaign may not be the next version, but it is certainly an intermediary one. With Trump, potential voters became not only consumers of political media, but also producers of it. His armies of supporters (and trolls) scoured the Internet, fighting comment board debates, posting catchy memes that delivered short-fire bursts of condensed political messaging, developing high-brow explanations of their positions and overwhelming the political opposition (many online publications deactivate the comment boards for certain articles, and the reason is not because of the spammers). His most fervent Internet loyalists are prosumers – consumers and producers of political ideas, who have become militant under the narrative of fighting against entrenched political and corporate interests. This outpouring of energy builds to the sort of critical mass where mass political affiliations shift and new platforms and networks gain in value, to the detriment of others.
NOTES