Just before Congress left Washington, D.C. for the August 2017 recess, Rep. Adam Schiff and Sen. Mark Warner — ranking Democrats on the House and Senate intelligence committees respectively — expressed interest in questioning representatives from big data firm Cambridge Analytica, a company that contracted with the campaign of President Donald Trump. Both the House and Senate intelligence committees are investigating whether the Trump campaign colluded with the Russian government to help him secure the presidency.
Spokespersons for Schiff, Warner, and Cambridge Analytica did not respond to our questions about whether the firm (which boasts on its web site that it uses data "to change audience behavior") has been contacted by either the Senate or House intelligence committees as Congress returns from recess in September.
Cambrdige Analytica and its British parent company, SCL Group, as well as President Trump's overall campaign digital operations which were overseen by his son-in-law and key adviser Jared Kushner, have been the subject of intensifying scrutiny from both lawmakers and the news media. Brad Parscale, the campaign's digital adviser, has agreed to be interviewed by the House Intelligence Committee. He tweeted a statement saying he was unaware of any Russian involvement.
On 25 July 2017, Schiff, a Democrat who represents the Los Angeles area, told "PBS NewsHour":
We’re looking at any of the Russian active measures that may have been employed here that we know they employ in other places. We’re looking at allegations concerning the social media campaign, whether there was any kind of cooperation or coordination through Cambridge Analytica.
Similarly, Warner, the vice-chair of the Senate Intelligence Committee, told CBS News' "Face the Nation" on 16 July 2017 that there was an undeniable presence of trolls on social media that may have influenced the U.S. elections:
[W]e do know that there was a series of Russian trolls, paid individuals, who worked for the Russian services that were trying to interfere and put fake news out. We also know they created what's called bots. In effect, internet robots that actually could interfere as well.
The question we have is: Did they somehow get information from some of the Trump campaign efforts to target that interference? We don't know that for sure. But what we do want to know is — I'd like to talk to the folks with Cambridge Analytica. I'd like to talk to some of the folks from the Trump digital campaign.
Writing for the Daily Beast, Clint Watts, a former Federal Bureau of Investigation counterterrorism agent and fellow at the Foreign Policy Research Institute who has been key in delineating Russia's active measures in the digital world, noted that regardless of whether Cambridge Analytica had any interactions with Russian influence agents in the 2016 election, they at least had overlapping goals (as in Brexit; the Kremlin has an interest in breaking apart the European Union). He wrote:
Some will claim that Russia might be seeking voter rolls to determine how to influence certain segments of the electorate in favor of Trump, but American social-media accounts openly provide far richer data on voters’ preferences and vulnerabilities for influence. Just ask Jared Kushner, his data-analytics wunderkinds, and their partner, Cambridge Analytica.
Cambridge Analytica’s primary backers, Robert and daughter Rebeka Mercer, simultaneously headed a pro-Trump PAC dubbed Make America Number 1, which was headed by White House counsel Kellyanne Conway. The PAC doled out $20 million in 2016 — and more to Cambridge Analytica than for any other expenditure. Trump White House Senior Adviser Steve Bannon sat on Cambridge Analytica’s board while running operations at Breitbart before becoming Trump’s campaign CEO.
Russia didn’t need to scout voter rolls to influence them for Trump on Facebook; Cambridge Analytica performed that targeting on its own, delivering stories based on hacked materials to Trump voters — doing the Kremlin’s work for them.
Experts say the way Cambridge Analytica and SCL Group use psychographics — targeted profiles of individual Internet users. But they've gained notoriety because of their involvement in recent consequential elections, like the 2016 U..S. presidential election and the Brexit referendum in which United Kingdom voters chose to leave the E.U.. The role of the company again came under scrutiny when it was hired by Kenyan President Uhuru Kenyatta, who was victorious as the incumbent in the 8 August 2017 general election, the aftermath of which saw deadly pre- and post-election violence.
Michal Kosinski, a psychologist and data scientist at Stanford University who is a pioneer in the field, said Cambridge Analytica's notoriety has less to do with what they do than how they go about it. Major political campaigns in the United States, no matter their affiliations, used the same tools — they just didn't talk about them. On the other hand, Cambridge Analytica were novices, and they "broke many laws, many rules" and spoke openly about their methods. Kosinski said. (Company representatives now deny using psychographics in the Trump campaign.)
Kosinski told us in a phone interview:
They learned the hard way and now they never mention they do it. They joined the ranks of people in the public relations world who don’t discuss how the sausage is made.
Jonathan Albright, research director for the Tow Center for Digital Journalism at Columbia University, said that the company used psychological profiles to appeal to voters' negative emotions, providing just enough of a nudge that they act on them in some way, even if that meant staying home instead of voting. One may argue political advertisements have long done just that -- political campaigns in the past would purchase ads that voters would see on television or in print en masse in key areas. But the big difference now is that people's digital devices tend to be connected via the Internet, so what they see online can follow them. And unlike standard political advertising, there is little regulation of what companies like Cambridge Analytica are doing in the United States.
Albright told us:
This is more an emotional pull versus push targeting where you're pushing [advertising] materials out to people. This grabs people through very specific location analysis and a bunch of different data points, to connect certain messages to real people at the right place and the right time, and probably when certain types of people are together.
If you visited a conspiracy web site at election time, that visit could follow you all the way back to your television and shopping purchases. You can get a specific type of television commercial based on a conspiracy web site visit or because you paused on something on Facebook.
So how would such a company tip an election? Albright told us:
For high-interest targets when you're dealing with a razor-thin margin to win a state, those [voters] are worth the extra time. They claim they have thousands of data points on every voter — I guarantee you for certain voters in Ohio or other swing states, the modeling was quite sophisticated.
In a Medium.com post, Albright explained how the company harvests data that allows them to target individuals with a new kind of precision:
CA has been called out for borderline ethical use of personal Facebook data, such as covertly gathering “likes” to predict the attitudes and beliefs that Facebook users might share unknowingly.
Other strategic information could include: connected third party application data; comments and likes on public Facebook pages; internet browsing history through Facebook APIs and scripts; consumer loyalty programs, mobile app logins; publicly shared photos and profile information that users forget about; and (I’m presuming) more mundane tactics such as harnessing unassuming personality “quizzes” on Facebook that capture invaluable psychometric data people readily share with their friends and families, but not with a psychological voter profiling firm.
Kosinski said the fact that Cambridge Analytica and its related firms have people fearful of the effects of big data is something of a moral panic; the methods in and of themselves are not inherently evil, and in fact are beneficial if used ethically. Data sharing is the crux of technological progress that defines the modern era. Without it, there would be no Google Maps, no Netflix, no self-driving cars, or even some advancements in medical sciences, among other things.
Although microtargeting of voters may raise paranoia about undermining democracy, Kosinski says that's somewhat misdirected:
People found a scapegoat to explain their defeat, because otherwise they would have to take the responsibility for not picking the right candidate or ignoring huge parts of the population. Instead they say, "It’s this modern technology used in this evil way to manipulate people." Moral panic sells much better than saying "Hey, stop for as second and think about what you're talking about."
In fact, he said, it can be good for democracy when used ethically, because it creates more plurality and involvement for a lower cost, and forces politicians to listen to what individual voters want:
If I were a politician and I just created one TV spot, I would need to come up with the broadest slogan, like "Yes we can," or, "Make America great again." But they don't really mean anything. Now, If I can talk with you in particular, which microtargeting allows, I'll need to talk to you about issues you care about. If I need to talk about issues, that means I, as a politician, start suddenly caring about what issues you care about. Being able to reach you and talk to you one-on-one suddenly makes politicians interested in inspiring a person and thinking hard about what this person wants and how to include it in their program.
For example coal miners — no one cared about coal miners, but Donald Trump found a way of reaching out to them. He did it with hundred s of other small groups that were broadly ignored, the same with Bernie Sanders. You can disagree with the solutions that they offered — I disagree with Trump’s proposals, but the answer to it is not try to find scapegoats in digital marketing, or say [Trump's voters] have been manipulated. The answer is to go out to those same people with the same tools, which are cheap and widely available, and say, "Hey, why not listen to me? I'll tell you why Trump is wrong."