Tech giants have “a whole class of techniques” exclusively at their disposal “for shifting people’s opinions, thinking, attitudes, beliefs, purchases, and votes without people knowing, and without leaving a paper trail,” Robert Epstein said in an interview with The Epoch Times for the “American Thought Leaders” program.
Epstein, a senior research psychologist at the American Institute for Behavioral Research and Technology and former editor-in-chief of Psychology Today, has devoted the past 6 1/2 years to researching tech giant bias, especially with Google, which dominates the search engine market.
“Americans see Google search results about 500 million times a day. Google controls roughly 90 percent of search. The next largest search engine, Bing, controls about 2 percent of search,” Epstein said.
Epstein’s peer-reviewed research found that research participants were remarkably susceptible to bias: search engine bias could easily shift 20 percent or more of the votes of undecided voters in an election. He also found that while results on Google leaned left substantially, results on Bing and Yahoo did not.
On Sept. 9, it was announced that 48 U.S. states, the District of Columbia, and Puerto Rico have opened a bipartisan antitrust probe into Google. The new investigation follows existing investigations by the Justice Department and the Federal Trade Commission into Facebook, Google, Apple, and Amazon.
One major form of bias featured in Epstein’s research is the search engine manipulation effect, something he began exploring after finding research in 2012 on how search result rankings affected purchases and clicks. It said that users tended to trust the highest-ranked search results the most, so much so that 50 percent of clicks went to the top two items.
To study how search engine rankings could shift voting preferences, he conducted a series of experiments in which he showed groups of randomly assigned people biased search results. They used a Google-like search engine, Kadoodle, that featured real search results and web pages taken from Google. The only difference was the ranking of the results.
One group was shown results biased to one candidate, one saw results biased to the other candidate, and the control group was shown mixed results, with bias in both directions. Before and after looking at the search results, participants were asked about their thoughts on the candidates, and who they’d vote for if they had to decide at that moment.
“I thought I could produce a shift in voting preferences and opinions of maybe 2 or 3 percent,” Epstein said. “The first experiment I ran, the shift I got was 48 percent.”
Epstein conducted more than a dozen different experiments, in which he found substantial shifts every time. In one large-scale national study across all 50 states with more than 2,000 participants, Epstein found that among different demographic groups, some were especially susceptible to manipulation, with shifts in preferences as high as 80 percent.
These shifts are reflective of actual behavior in the ballot box, Epstein said. Survey research has shown that “if you ask people who they’re going to vote for, it turns out that’s a very good predictor of who they actually vote for,” he said. “Generally speaking, we’re talking about 90, 95 percent accuracy in predictions.”
And according to Epstein, what they found in their experiments likely underestimated the real impact that Google has since most of his experiments had participants conducting only one online search.
“In real life, people are conducting many searches over a period of weeks or months that are election-related. If they’re undecided, that means they’re being hit over and over and over again with biased search results, taking them to web pages that favor one candidate,” he said.
So far, Epstein has identified 12 major techniques of tech giants that can shift perceptions and opinions.
Another easy way to influence voters is by targeted messaging, such as sending out “go vote” reminders only to people with certain political biases. Based on Epstein’s calculations using Facebook’s published data in 2012, Hillary Clinton would have received 450,000 more votes on election day in 2016 if Facebook had sent a “go vote” reminder just to left-leaning users.
No Paper TrailIn early 2018, a leak to The Wall Street Journal included one email from a Google employee that mentioned using “ephemeral experiences” to counter Trump’s immigration policy.
“What’s an ephemeral experience? That means you type in something, let’s say a search term. And some results are generated on the fly just for you. They impact you, they disappear; they’re gone. And they’re not stored anywhere. And you can’t go back in time and find them,” Epstein said.
“This is a fantastic way to manipulate people,” because it leaves no paper trail to follow, and people rarely spot the bias, he said.
“And here’s something creepy. The very, very small number of people who can spot the bias, they shift even farther in the direction of the bias.
“Most of these types of influence have never existed before in human history. They’re made possible by the Internet. They’re made possible by these huge tech monopolies, and they’re entirely in the hands of these tech monopolies.
“In elections, we’re influenced by billboards, by radio shows, and TV shows, and advertisements, and so on. All of that is competitive. And in that sense, it’s probably a good thing. It’s a good thing for democracy that there is so much competition out there vying for your attention and trying to convince you of this or that. But if there’s bias in search results, that’s controlled by the platform, in this case, Google. That’s not competitive.”
2016 ElectionIn 2016, Epstein set up a secret monitoring system that showed that Google results were significantly skewed toward Clinton in the months leading up to the presidential election.
Epstein had 95 field agents in 24 states conduct election-related searches with neutral search terms on Google, Bing, and Yahoo. The results from all these searches were then saved.
“We were able to preserve 13,207 election-related searches as well as the 98,044 webpages to which the search results linked,” Epstein said. In effect, they were able to permanently preserve snapshots of what are normally “ephemeral” experiences.
Epstein decided only to collect the data, but not to analyze it prior to the 2016 election, because if he found bias, he would face an impossible dilemma.
“What would I do? I mean, if I announced it, there would have been absolute chaos, especially, I think, if there was bias against Donald Trump. And if I didn’t announce it, then I would be complicit in the rigging of an election,” he said.
In the analysis, “we found substantial bias favoring Hillary Clinton in all 10 search positions on the first page of search results on Google, but not Bing or Yahoo,” he said, adding that the probability that the bias was solely due to chance was less than 1 in 1000.
Through a series of calculations, Epstein concluded that if this level of bias was present nationwide, it would’ve shifted somewhere between 2.6 million and 10.4 million votes to Clinton.
Epstein describes himself as a moderate who leans liberal. And he had been a longtime supporter of the Clintons. “But I felt very strongly that since our results were so clear that I had a responsibility to report the findings,” he said.
Clinton won the popular vote by more than 2.8 million votes, but the popular vote “might have been very different,” Epstein said, if there had been no bias in Google’s search results.
“It was uncomfortable for me to have to acknowledge that, to have to announce that. But that’s what I concluded from the research.”
People trust in Google’s search rankings, he said, because they believe it’s generated by a computer algorithm, and thus must be impartial. What was especially disturbing was the subliminal manipulation; in most cases, “people can’t see the bias in search results.”
For the 2018 midterm elections, Epstein set up a larger monitoring system focusing on three Republican districts in Orange County, California, which all ended up flipping Democrat. He found that on Google (but not Bing or Yahoo), search results were strongly biased in favor of Democratic candidates.
Is Tech Giant Bias Intentional?Google has insisted that their algorithms for search ranking evolve according to the “organic” activity of users interacting with the algorithm.
“In my mind, that’s complete nonsense. I’ve been a programmer since I was a teenager,” Epstein said. “The fact is, Google has total control over what happens.
“Let’s say there are a lot of users who lean left or who lean right, the algorithm can respond any way it’s programmed to respond. So I simply don’t buy the idea that this was just the algorithm’s fault or just the user’s fault.”
Although he believes Google has total control of its search ranking, that doesn’t necessarily mean Google engineers deliberately designed their algorithms to have left-leaning biases.
“I admit I used to be obsessed with wanting to know whether executives at a company like Google or just rogue employees were fiddling around with search results and search suggestions,” he said.
But now Epstein doesn’t think it matters. Regardless of whether people working at the tech giants have intentionally skewed the results, or if they’ve simply been negligent about political bias, the reality is that they have an enormous impact on thinking, behavior, and votes even in distant corners of the world, he said.
“Let’s say in many countries, they don’t care. For many elections, let’s say they don’t care, but the algorithm is still going to do its thing,” he said. “Their algorithm is meant to tell you what’s best, and what’s best goes at the top.
“So what I realized was it’s very possible that a lot of important events right now in human history are being determined not by plans and goals and strategies of human beings at a company like Google, but by computer programs that are just being left to do their own thing. To me, that’s far more frightening than thinking that a Google executive is out to rule the world.
Combating Bias in 2020For the 2020 elections, Epstein plans to launch a much more ambitious monitoring system to track tech-giant bias.
“I think that the tech companies are going to go all out” in 2020, he said. “I think they were very cautious and overconfident in 2016. I think there’s a lot of crazy things they could have done to shift votes that they just didn’t do.”
As the 2020 election nears, Epstein plans to set up at least 1,000 field agents in all 50 states. “And we’re planning this time to use artificial intelligence—we’ve been working on this in recent months—to analyze the massive amount of data that we’re receiving every day in real time. This means that if we find evidence of bias or some sort of manipulation, we‘ll announce it. We’ll announce it as soon as we’re sure that we found it ... either to the media or to the Federal Election Commission or to other authorities,” he said.
“And that is going to create a kind of chaos. But it’s the kind of chaos we need to have.” He hopes that this monitoring project will encourage tech giants to stop political bias on their platforms.
On the other hand, he said, “if they don’t back down, and we continue to detect and capture evidence of large-scale vote manipulation, I think frankly these companies will pay a terrible price. I think there could be both civil actions taken against them and possibly criminal actions taken against them.”
Either way, Epstein said, “democracy wins. And that’s my concern here. I’m not concerned about any particular party or candidate, although I do lean left. I’m concerned about democracy and free and fair elections.”
He believes that Google could easily remove political bias in its search results using techniques it’s already developed to deal with what they describe as “algorithmic unfairness.”
Such techniques were thrust into the spotlight by the massive trove of documents recently leaked by former senior Google software engineer Zachary Vorhies. A simple example is the search term “American inventors.” Whereas the original results might have shown a majority of white males, more black Americans can be boosted to the top of search results to make the results more “fair.”
If machine learning fairness techniques can correct for what Google engineers see as racial unfairness, the same could easily be done for political bias, in Epstein’s view.
“And I think we have to think beyond the United States because a company like Google is impacting more than 2 billion people around the world. Within three years, that number will swell to over 4 billion people,” he said.
“They can literally impact thinking behavior, attitudes, beliefs, elections in almost every country in the world.
“In my mind, that means building larger, better-monitoring systems to keep an eye on companies like Google. I think that’s necessary, not only to protect democracy around the world but to protect human autonomy.”