Who doesn't search, doesn't find
but who searches, might not find either
Election season had arrived here at the ranch. We had even allowed some politicians to put up signs visible from the road. Cowboys Jack and Pete were doing some fence repairs around the biggest of all of these signs, which was a sign for the nation’s foremost conservative contender, Pierre-Paul Delièvre. “Back in the day,” Jake said, “we’d always say: ‘Never vote for the French guy’. But this time around, I’m gonna change my mind. Delièvre is saying some great stuff that’ll bring some common sense back.” “Sure he will?” replied Pete. “I only find LookingGlass links to bad stuff about him.” “That can’t be true,” stammered Jake.
Sure, Pete was right. His LookingGlass search results looked as follows:
What neither Jake nor Pete knew though, was that LookingGlass had preselected the winning candidate, Absynthia Freilandt, long before they had thought of doing any research. In real life, Delièvre had made some very reasonable statements, such as pleading for a zero tolerance for violent aggression of any kind, without lenience for people with certain “characteristics,” whereas Freilandt had done the opposite. She even advocated for cruelty like closing the bank accounts of people who “undermine trust in the democratic institutions.” “Those deplorables will have to eat dust,” she had once infamously stated. That did not deter LookingGlass, though. If Freilandt won, LookingGlass would deliver the architecture and solutions the government would need to identify those citizens who “undermine trust,” so it had a stake in assuring Freilandt’s victory, who could then turn the country into the opposite of her name. Could LookingGlass’s efforts ever have a decisive impact on the election outcome, though?
The answer to the above question is a simple and affirmative: yes, it can. Solid research results have established the existence of a search engine manipulation effect (SEME) on the outcome of elections. The SEME allows dominant internet search providers to sway the opinions of undecided voters in the direction of their candidate of choice. Here is what the authors of the seminal paper on SEME conclude:
“The results of these experiments demonstrate 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) search ranking bias can be masked so that people show no awareness of the manipulation. We call this type of influence,which might be applicable to a variety of attitudes and beliefs, the search engine manipulation effect. Given that many elections are won by small margins, our results suggest that a search engine company has the power to influence the results of a substantial number of elections with impunity. The impact of such manipulations would be especially large in countries dominated by a single search engine company.”
The authors arrived at these conclusions by conducting three clever experiments in which undecided voters were asked to only do internet research on either of the candidates in Kadoodle, the mock search engine set up by the organizers of the trial. Kadoodle would show certain cohorts of participants search results biased in favour of one or the other candidate. By polling the participants’ opinions, it thus became possible to quantify the effects of search engine manipulation.
(I wish search engine manipulation pointed readers to Wild Horse Wisdom, but as long as it does not, I’d appreciate if you could subscribe!)
While the results from the individual experiments differed, the authors concluded that the opinions of one in five of undecided voters can be nudged in the direction of the bias in the search results. Doing a back-of-the-envelope calculation, we find the following. Let’s assume that 90% of voters use a given search engine, and that 10% of those 90% are undecided. If search engine manipulation can sway 20% of the group of undecided voters, it can manipulate the election outcome by a margin of 1.8% each way, which is more than the margin in many closely contested elections. In fact, per the authors, 25% of national elections around the globe are decided by margins under 3%. The authors therefore conclude their paper with the ominous statement
“We conjecture, therefore, that unregulated election-related search rankings could pose a significant threat to the democratic system of government.”
I would argue that SEME by itself is a much graver threat to democracy than any of the candidates or fringe movements in any Western country. In the “West,” Google is the dominant search engine. According to most estimates, its market share exceeds 90%. Recently, Google even lost an antitrust lawsuit is which their practices were ruled to be “monopolistic.” Being the dominant provider of internet search, Google is in a position to unduly influence any election won by a narrow margin, such as, say, any of the US swing states. That would not be an issue if US-based big techs had any ethics, but unfortunately, they blatantly abuse their position. They do so through SEME, but more recently also through other effects. For instance, Google has banners above its search bar that can appear, such as a reindeer and a Christmas tree in the holiday season. However, by selectively showing such banners about ‘Election Day’ to certain users, in addition to swing voters’ opinions, turnout by a given voter cohort can be influenced too.
The Search Suggestion Effect (SSE) is a similar effect generated by auto-complete: when the user starts typing, Google will come up with automated suggestions that clearly expose its political bias. For instance, when I start to type “why censorship is …” I get the following suggestions:
Just one of the above suggestions questions the need for censorship. But that merely results from objective ranking, right? Well, let’s have a look at the same search autocomplete results from DuckDuckGo, a less popular engine.
DuckDuckGo presents exactly four positive and four negative auto-completes, with the negative ones more dominant at the top of the list. In spite of the more balanced results, even in DuckDuckGo’s list some obvious answers do not appear, such as “totalitarian,” “incompatible with democracy” or “a common tool for repression.” What is truly pernicious about these nudging activities is that they are set up in a way that psychologically abuses the user’s trust: by presenting trustworthy results to apolitical questions, users are being manipulated into thinking that search results are unbiased. When those same users then research a more political question, they will likely still trust the results presented, in spite of their obvious bias. And of course, nudging is not limited to search. We can also observe social media manipulation effects, as well as legacy media manipulation effects. The latter are well documented in the case of Fox News in the United States.
Elections around the world have been interfered with by SEME and other digital nudging effects in a grotesque way. Moreover, the interference is very hard to prove, since different users see different search results and social media profiles. Election integrity officers may not see a banner that reminds them to go vote when they open their search page, whereas users with a profile friendly to the preferred candidate may well.
The above results are indicative of Google’s general trend in nudging, which always seems to boost opinions and politicians who stand for control and oppression, and are antagonistic to individual rights, spirituality, and enlightenment. It would not surprise anyone to see such an attitude from a company that reports to a fascist regime, such as Baidu, but why are we seeing this from a company that claims to represent the values of the West?
(If it seems contradictory that a self-proclaimed communist country is fascist, please check out an earlier exposé on such directional dysphoria and consider to subscribe)
At this point, one can only conjecture, but the trend is clear. SEME and other nudging effects seem to be consistently applied to favour politicians who classify as “LINOs.” That is, “liberals” in name only, who suffer from severe vocabulary distortion induced directional dysphoria, who run on vacuous campaign programs and slogans, yet once elected, do not hesitate a second to take illiberal measures, such as to restrict speech and to silence opponents. Two very good examples are observed in Brazil and, more recently, the UK, where the so-called “left-wing” Labour Party has not even needed a month to abuse a horrific stabbing to propel an agenda of increasing limits to individual freedom. Labour could have reacted to the events by detailing a list of actions to prevent further stabbings or by addressing the root cause of the riots. Instead, they diverted attention to “harms caused by online anonimity” and have drastically reduced freedom in the public discourse by arresting people who expressed emotional reactions to the events for “incitement to violence” (just one example can be found here). Illegal incitement to violence means, in any country, a statement that urges to action. To merely conjecture that the perpetrator of the stabbings adhered to a certain religion, ethnicity or immigration status, is a theory and by no means incites violence. The recipients of that information still have a responsibility to independently verify what they read and hopefully, have their own moral compass that should keep them away from aggression. Yet arrests were made for “incitement to violence” based opinions alone. Regardless, Labour politicians, as well as the Brazilian Supreme Court, continue to be portrayed positively by Google, even after they started to emulate Soviet practices.
It seems though, that Google’s endgame is tightly inosculated with the self-proclaimed “liberal” politicians’ strategies: by using insidious online manipulation effects and vacuous campaign slogans, liberals in name only want to establish “what can be” a single-party totalitarian society that proffers its elites unchecked corruption, as it is “unburdened by what have been” the nuisances of human rights, due process, objective science and the separation of powers.
We need not walk along that path. To insulate ourselves from such unacceptable outcomes, we should pass legislation that mandates online search providers to open source their algorithms. That way, developers can tweak them to provide objective results, or results with a certain filter to be selected by the user, such as a conservative or liberal filter. Users will then be at liberty to explore opposing viewpoints and distill their own synopses. While we don’t have access to the algorithms, we should take any effort we can to depend as little as possible on the dominant market players. That does not only apply to search, but for instance also to e-mail. There are several independent mail providers who respect privacy and are, from a user’s perspective, offering services on par with Google.
The build-up into a totalitarian state may be frightening, but we can rest assured that if we keep faith in ourselves, it will not succeed. We horses have seen many tyrants come and go. The herd has always stuck together. Our leader was always the most capable horse. We did not need to be nudged into respecting our leader. Nor have leaders ever needed to change their gender to earn respect.
Our cowboys are out in the open most of the day. They don’t spend hours on internet search. When Jake and Pete returned, Pete said: “You know, Jake, when I chatted with the crew at Five Mile, they all agree that the French guy is reasonable. I’ll just ignore the West Coast horseshit.” In the end, reason wins as long as people have a connection with reality. In our parts of the world, there is no need to create horseshit. It just drops to the floor.
(To the interested reader: I have recently been posting short comments and preview snippets on X. A warm welcome to every reader who joins the herd there!)
Freedom of speech must remain top priority, even, and especially, we disagree. "Online hate" and "incitement" are just another mechanism to silence those they don't want to be heard. It's only incitement when they don't like what you say. It's definitely not incitement if you're
calling to explicit action for a cause they approve of. Manipulating search algorithms adds to that. If a comment is posted but nobody reads it, does it even make a sound? Freedom of information and freedom of speech go hand in hand, and we neet to fight for both.