Algorithms & Elections
How Google & Facebook Can Manipulate Election Results
The idea of election manipulation has been circulating in the media for a few years now. Their focus seems to be aimed at nefarious foreign actors trying to influence a domestic vote. While this is a problem, there appears to be something missing.
If foreign actors can manipulate local elections, what about domestic actors? In particular, what about the organizations that control the information we see every day? If a Google or Facebook wanted to manipulate an election result or push voters in a certain direction, couldn’t that be more damaging than foreign intervention?
Our focus on manipulation from abroad is well founded, however there are possible nefarious actors right within our borders who should be examined as well. According to a 2018 article in Business Insider, Google has a 90% market share on U.S. internet searches. A Pew research study found nearly 50% of Americans get news through Facebook.
What if these gatekeepers of our data and social media decided to manipulate us to vote one way or another? A good deal of research has been done that shows exactly how much Google and Facebook could turn a voter’s will. According to at least one data scientist, this is their plan for the present and future.
Gerrymander: (verb) manipulate the boundaries of (an electoral constituency) so as to favor one party or class.
— Definition provided by Google
In an article in the Harvard Law Review, Jonathan Zittrain explained an experiment conducted by Facebook on November 2, 2010. Facebook wanted to see if they had an ability to move their users to vote. They conducted this experiment on 60 million users during the 2010 congressional elections.
Users were shown an image asking them if they planned to vote. This image also showed other Facebook friends who had voted. The image had information for local polling stations as well. Facebook compared users who were shown this image to a control group who were not. The users in the experiment were cross referenced against names who voted at the local polling stations to see who actually voted.
Facebook found that the users shown the image were .39% more likely to vote than those who weren’t shown the image. It sounds like the effort was insignificant, but that percentage equated to 60,000 voters.
Furthermore, Facebook found there was a “ripple effect”. In this ripple effect, close friends of the users who saw the image were more likely to vote. This ripple effect worked out to be 340,000 people who were influenced to vote. A number like this, even across the whole United States, could change election results. This is especially valid when you consider there were a number of congressional elections that were determined by thousands of votes.
Jonathan Zittrain poses an interesting question after looking at this experiment. What if Facebook decided to go through with this experiment for real to influence an election? What if they chose to show this vote message to users who identified with one political party? That would more than likely enable the social media giant to manipulate an election. It would also barely be noticed.
Zittrain referred to a possible manipulation like this as “digital gerrymandering”. It would enable one CEO or group of high level corporate executives to exercise an incredible amount of influence over an election without being noticed. At least with monetary donations, they flow both ways. A cyber gatekeeper this size wouldn’t be challenged.
Google’s Power To Manipulate
I am Dr. Robert Epstein…Senior Research Psychologist at the American Institute for Behavioral Research and Technology. I love America and democracy…. I am here today for three reasons: to explain why Google presents a serious threat to democracy and human autonomy, to explain how passive monitoring systems can protect us both now and in the future from companies like Google, and to tell you how Congress can immediately end Google’s worldwide monopoly on search… I have been a research psychologist for nearly 40 years and have also served in various editorial positions at Psychology Today magazine and Scientific American MIND. I received my Ph.D. at Harvard University in 1981 and have since published 15 books and more than 300 scientific and mainstream articles on artificial intelligence and other topics. Since 2012, some of my research and writings have focused on Google LLC, specifically on the company’s power to suppress content — the censorship problem, if you will — as well as on the massive surveillance the company conducts, and also on the company’s unprecedented ability to manipulate the thoughts and behavior of more than 2.5 billion people worldwide. Data I’ve collected since 2016 show that Google displays content to the American public that is biased in favor on one political party (Epstein & Williams, 2019) — a party I happen to like, but that’s irrelevant. No private company should have either the right or the power to manipulate large populations without their knowledge. — Dr. Robert Epstein’s Testimony To Congress June 16, 2019
Dr. Robert Epstein, a research scientist with 40 years of experience in his field, has been ringing a warning bell about Google for a few years now. He believes that Google can manipulate people not by content, but by what content is chosen to be shown to searchers. He also believes that how the content is ordered can also manipulate users.
In an article in the Epoch Times, Epstein explains how he conducted peer-reviewed, randomized, and controlled studies done with thousands. In these studies, Epstein could repeatedly shift undecided people’s opinions on most topics.
He worked this magic just by filtering information that got shown to the subjects in multiple searches. The creepy thing about the experiment was that the subjects couldn’t tell they were being manipulated. He identified 10 ways Google can manipulate voters, but 2 of the methods were particularly powerful.
- ) Search Engine Manipulation Effect (SEME): In this manipulation favored results are listed first in a search result. His test subjects believed that the items listed first in the search held the best information. By this manipulation, Epstein was able to shift the preferences of undecided subjects by 20% — up to 80% in certain demographics.
- )Search Engine Suggestion Effect (SSE): This manipulation involves the auto-fill feature in searches. You’ll notice this when you begin to type something in the Google search and it completes the statement for you. Epstein in his research has shown by using this manipulation he can turn a 50/50 split on an issue to a 90/10.
It’s Just The Algorithm Doing What Algorithms Do
Google’s response to the accusation that they manipulate algorithms is to claim they do no such thing. One of their defenses is that the algorithms change themselves, sometimes in unpredictable ways. They bend and shape to their internal programming and the users they interact with.
There have been instances where algorithms went out of control in ways their creators could never have predicted. In an article in TechCrunch, Megan Rose Dickey examines some instances of out of control algorithms.
In one instance, a Microsoft AI bot named Tay turned into a white supremacist. In another instance a Massachusetts Registry of Motor Vehicles facial recognition algorithm labeled an innocent person a criminal and revoked their driver’s license. In another event, Google searches when asked if the holocaust really happened brought up a Nazi friendly site as the first result.
Dickey in her article makes the point that people are rarely held responsible for the act of an algorithm. The article also mentions that these algorithms function as a sort of ‘black box’. The tech giants that use them don’t allow the public to look inside. A Google and Facebook may claim innocence for the action of an algorithm, but how could one ever know.
Dr. Epstein doesn’t believe Google’s claim of innocence. He refers to an internal video that was leaked to the website The Verge. The video created by Google called “The Selfish Ledger” explains how data collection has the ability to reshape the world. The creepy video explains how data can be used to change people’s behavior the same way altering DNA can change the human body.
“Whilst the notion of a “global good” is problematic, topics would
likely focus on health or environmental impact to reflect Google’s values as an organization… Over time, by selecting these options, the user’s behavior may be modified, and the ledger moves closer to its target…As this line of thinking accelerates and the notion of a goal-driven ledger becomes more palatable, suggestions may be converted not by the user, but by the ledger itself…Our ability to interpret user data, combined with the exponential growth in sensor enabled objects, will result in an increasingly detailed account of who we are as people. As these streams of information are brought together, the effect is multiplied. New patterns become apparent, and new predictions become possible…. Just as the examination of protein structures paved the way to genetic sequencing, the mass multigenerational examination of actions and results could introduce a model of behavioral sequencing…As patterns begin to emerge in the behavioral sequences, they too may be targeted. The ledger could be given a focus, shifting it from a system which not only tracks our behavior but offers direction towards a desired result.” — Excerpts from Google’s “Selfish Ledger” video
What Does The Future Hold?
Epstein in his article in the Epoch Times explains that Europe has already started cracking down on Google. Google was recently fined $2.7 billion dollars for unfairly demoting competitors in their search results. The European Commission also fined Google €4.34 billion for antitrust violations.
Domestic pressure has begun as well on the two digital gatekeepers. Facebook was recently fined 5 billion dollars for misuse of user data. Peter Thiel, a board member of Facebook, publicly accused Google of treason for their work with the Chinese government.
“Google refused to work for the Pentagon on artificial intelligence [AI]. If you turn around and you work on AI in China, and you don’t really know what they’re going to do with that, I think there’s an issue.”
— Richard Clarke, Obama-era White House cybersecurity chief
It would appear the future for the two tech gatekeepers will be one in which they’re much better monitored on various fronts. As for the election manipulation front, Dr. Epstein himself is taking action. He’s working with a number of colleagues to create a system to monitor data manipulation by the tech giants.
“As I mentioned earlier, in 2016, I led a team that monitored search results on multiple search engines. That was a start, but we can do much better. These days, I’m working with business associates and academic colleagues on three continents to scale up systems to monitor a wide range of information the Big Tech companies are sharing with their users — even the spoken answers provided by personal assistants. Ultimately, a worldwide ecology of passive monitoring systems will make these companies accountable to the public, with information bias and online manipulation detectable in real time.
With November drawing near, there is obviously some urgency here. At this writing, it’s not clear whether we will be fully operational in time to monitor the midterm elections, but we’re determined to be ready for 2020.” — excerpt from Dr. Epstein’s article in the Epoch Times.
Thank you for reading my ramblings. If you’ve enjoyed what you’ve read, please share.