From search engine algorithms to electronic voting machines, technology provides opportunities for manipulation of voters and their votes in ways that could profoundly affect the results of the 2016 election. In the US, the 2012 presidential election was won by a margin of just 3.9 percent; and, historically, half of US presidential elections have been won by margins under 7.6 percent. These narrow but consequential victory margins underscore the importance of understanding how secret, proprietary technologies—whether they are newly developing or increasingly outdated—potentially swing election results.
Mark Frary, in Index on Censorship, describes the latest research by Robert Epstein and Ronald E. Robertson of the American Institute for Behavioral Research and Technology on what they call the Search Engine Manipulation Effect (SEME). Their research focuses on the powerful role played by the secret algorithms (including Google’s PageRank and Facebook’s EdgeRank) that determine the contents of our Internet search results and social media news feeds.
Epstein and Robertson studied over 4,500 undecided voters in the US and India, using randomized, controlled, double-blind methods, with research subjects who matched as closely as possible each country’s electorate. “The results,” Frary reported, “were shocking.” Epstein and Robertson showed that biased search rankings “could shift the voting preferences of undecided voters by 20% or more.” The effect could be greater than 20 percent in some demographic groups, and—perhaps most significantly—this search-ranking bias “could be masked so that people show no awareness of the manipulation.”
In an earlier article for Politico, Epstein wrote that the Search Engine Manipulation Effect “turns out to be one of the largest behavioral effects ever discovered … We believe SEME is a serious threat to the democratic system of government.”
Epstein described how the study’s measures—including research subjects’ trust, liking, and voting preferences—“all shifted predictably” based on information provided by a Google-like search engine that he and Robertson created, which they called Kadoodle. In one of the experiments, Epstein and Robertson documented SEME with real voters during an actual election campaign: In a study involving 2,000 eligible undecided voters in India’s 2014 Lok Sabha election, they found that “search engine rankings could boost the proportion of people favoring any candidate by more than 20 percent—more than 60 percent in some demographic groups.”
Predictably, Google challenged these findings. As Frary reported, a senior vice president at Google, Amit Singhal, responded in Politico, “There is absolutely no truth to Epstein’s hypothesis that Google could work secretly to influence election outcomes. Google has never ever re-ranked search results on any topic (including elections) to manipulate user sentiment.” However, as Frary duly noted, “Singhal specifically says ‘re-ranked’ rather than ‘ranked.’ What he means by this is that the algorithm decides on the ranking of search results and that no one goes in and manipulates them afterwards. Google’s stated mission to ‘organize the world’s information and make it universally accessible and useful’ should perhaps have a caveat—‘as long as our algorithm decides you should see it.’”
Hidden algorithms shape online content in significantly different ways from more widely recognized concerns about editorial censorship on television and in print. On TV and in print, Frary observed, “there is a person at the heart of the decision process … We can imagine how commissioning editors think, but the algorithms behind Facebook and Google are opaque.” This concern has led Emily Bell, a journalism professor at Columbia University, to observe, “If there is a free press, journalists are no longer in charge of it. Engineers who rarely think about journalism or cultural impact or democratic responsibility are making decisions every day that shape how news is created and disseminated.”
When filtering is financially motivated, secret, and beyond our control, Robert Epstein told Index on Censorship, “we should be extremely concerned.” Online filtering on massive platforms such as Google and Facebook, he warned, is “rapidly becoming the most powerful form of mind control that has ever existed.”
More than 75 percent of online searches in the US are conducted on Google—in other countries Google’s share of Internet searches is as high as 90 percent; some 1.5 billion individuals, political parties, businesses, and other organizations now use Facebook. Epstein and Robertson are now researching how to counter SEME. “We found the monster; now we’re trying to figure out how to kill it,” Epstein wrote in his Politico article. These efforts hinge in part on eroding public trust in Google, including our willingness to accept whatever our search results present to us as fact.
As Frary reported, Facebook, Google, and others are “highly secretive about how their algorithms work.” Electronic voting machines present similar challenges, as Harvey Wasserman and Bob Fitrakis document in their book, The Strip & Flip Selection of 2016: Five Jim Crows & Electronic Election Theft. “Electronic voting machines are owned by private corporations … And the courts have ruled that the source code on these electronic voting machines is proprietary,” Wasserman told Amy Goodman of Democracy Now! in February 2016.
In 2016, about 80 percent of the US electorate will vote using outdated electronic voting machines that rely on proprietary software from private corporations, according to a September 2015 study by the Brennan Center for Justice at the New York University School of Law. Forty-three states are using machines that will be at least ten years old in 2016; in fourteen states, machines will be fifteen or more years old. The Brennan Center study identified “increased failures and crashes, which can lead to long lines and lost votes” as the “biggest risk” of outdated voting equipment, while noting that older machines also have “serious security and reliability flaws that are unacceptable today.”
“From a security perspective,” Jeremy Epstein of the National Science Foundation noted, “old software is riskier, because new methods of attack are constantly being developed, and older software is likely to be vulnerable.” Virginia recently decertified an electronic voting system used in twenty-four of its precincts after finding that an external party could access the machine’s wireless features to “record voting data or inject malicious data”. The investigation also raised concerns over the AccuVote-TSx machine, which is used in over twenty states. In 2014, voters in Virginia Beach observed that when they selected one candidate, the machine would register their selection for a different candidate, due to an “alignment problem.”
Amy Goodman of Democracy Now! asked Wasserman how voters using electronic voting machines could be sure that their votes are counted. He told her, “They can’t be. You cannot verify an electronic voting machine … The proprietary software prevents the public from getting access to the actual vote count.” In a March 2016 article on the Free Press website, Fitrakis and Wasserman wrote that the “veracity of outcomes” in electoral races for the offices of president, US Congress, governorships, state legislatures, county commissioners, and others “will vary from state to state based on the whims and interest of those in charge of the electronic tallies.”
On Democracy Now! and elsewhere, Wasserman and Fitrakis have advocated universal, hand-counted paper ballots and automatic voter registration as part of their “Ohio Plan” to prevent stripping and flipping in US elections.
Corporate media outlets including CNNMoney, Fortune, and the Washington Post provided some coverage of Epstein and Robertson’s research. In May 2016, the Huffington Post published an article by actor and activist Tim Robbins, titled “We Need to Fix Our Broken Election System.” “Every broken machine, every disenfranchised voter, every discrepancy between the exit polls and the final results,” Robbins wrote, suggests “malfeasance” and “leads to more and more disillusionment that results in less and less voters.”
Robert Epstein, “How Google Could Rig the 2016 Election,” Politico, August 19, 2015 http://www.politico.com/magazine/story/2015/08/how-google-could-rig-the-2016-election-121548.
Mark Frary, “Whose World are You Watching? The Secret Algorithms Controlling the News We See,” Index on Censorship 44, no. 4 (December 2015), 69–73. (Extract available via: http://ioc.sagepub.com/content/44/4/69.extract)
Lawrence Norden and Christopher Famighetti, “America’s Voting Machines at Risk,” Brennan Center for Justice (New York University School of Law), September 15, 2015, https://www.brennancenter.org/publication/americas-voting-machines-risk.
Harvey Wasserman, interview by Amy Goodman, “Could the 2016 Election be Stolen with Help from Electronic Voting Machines?” Democracy Now!, broadcast February 23, 2016, transcript, http://www.democracynow.org/2016/2/23/could_the_2016_election_be_stolen.
Bob Fitrakis and Harvey Wasserman, “Is the 2016 Election Already being Stripped & Flipped?,” Free Press, March 31, 2016, http://freepress.org/article/2016-election-already-being-stripped-flipped.
Student Researchers: Brandy Miceli (San Francisco State University) and Amanda Woodward (University of Vermont)
Faculty Evaluators: Kenn Burrows (San Francisco State University) and Rob Williams (University of Vermont)