In 2014, a survey app designed by Aleksandr Kogan collected data on 50 million Facebook profiles. Less than 300,000 of them had consented to this collection.
In 2015, he gave that data to Cambridge Analytica, a voter-profiling company. They used 30 million of those profiles to construct psychographic profiles of voters.
In 2016, the Trump campaign hired Cambridge Analytica to work on the campaign.
And now, in 2018, everyone’s angry about it. Mostly they’re angry at Cambridge Analytica. Phrases like “illegal harvesting of data” and “grossly unethical experiment” abound. But Kogan, Facebook, and the Trump campaign are taking fire too.
This is a story about data privacy, and changing attitudes toward it. But who’s the bad guy here? What are people really angry about? What can be done? And, in the end, are we focusing on the wrong thing? To find out, we need to start back in 2012.
2012: Privacy Concerns? No Big Deal
Big data and social media didn’t arrive on the political scene in 2012. But the scale on which they were deployed was groundbreaking. Data—especially data from social media—played a huge role in Barack Obama’s re-election campaign.
Facebook has provided tools for political campaigns for years. But the Obama campaign went beyond that. Canvassers, organizers, and other Obama supporters “consciously or otherwise” handed over public information from their Facebook profile.
Combined with all of the other ways an organization can buy people’s data, the campaign was able to put together comprehensive profiles of voters. Those profiles were used in ad targeting.
Here’s Obama’s campaign manager, Jim Messina, talking about the importance of big data in the election (don’t miss the part where he talks about running 62,000 election simulations every day):
This type of targeting was easier to do online than on TV, according to Ed Pilkington and Amanda Michel:
“[H]e said [addressable advertising] was unlikely to happen in any great quantity in 2012 as there are too many hurdles, including concerns in Washington about the privacy of cable TV consumers.
No such impediment will hold back the digital explosion this year.”
In 2012, concerns about privacy were viewed as an impediment to advertising—but only on TV.
To be fair, some privacy experts did warn citizens about the centralized databases used by the campaign. But in general, the press coverage of the campaign’s efforts to use big data, including a huge amount of Facebook data, were positive. There were some dissenting voices, but not as many as you’d expect in today’s post-Snowden world.
In 2012, we saw just how valuable social data is to political campaigns. We should have known someone like Cambridge Analytica would be coming.
How Does Facebook Data Influence Elections?
The connection between Facebook data and elections isn’t intuitively clear. What good does knowing if someone likes the movie Frozen do for a political campaign?
One of the most important pieces of information that campaigns get is the identity of influencers. Algorithms see which individuals are influential among their social groups, and those people are targeted for advertising. Sway the influencer, the thought goes, and they’ll sway their friends.
Much of the data collected also serves to target political ads. Facebook data can be sliced by geographic region, age, gender, interests, likelihood to vote for a specific candidate, and more. Campaigns can use information gathered on specific demographics to better target their ads.
Here’s how microtargeting works in political advertising:
Highly targeted ads can follow individual users around the internet, transmitting specific messages that campaigns believe are likely to swing them to their side. By using information collected from Facebook and third parties, campaigns can target users with alarmingly specific messages.
And these types of campaigns have been very successful. (Though not without cost. Chuck Todd makes an interesting argument that big data broke American politics.) With the amount of money changing hands in the political industry, there’s no question that this type of analysis would be highly sought-after.
2018: The Outcry Over Facebook Data Harvesting
Early in 2018, Cambridge Analytica hit the news in a big way. It’s a voter-profiling company that was hired by the Trump campaign during the 2016 election.
To make a long story short, Cambridge Analytica used Facebook data on millions of users that hadn’t consented to collection or analysis:
Note: Interestingly, the Guardian first brought allegations of suspicious data collection to light back in 2015.
Who Is Cambridge Analytica?
Cambridge Analytica is a voter-profiling company that’s owned by Strategic Communications Laboratories (SCL), a group that claims to have “conducted behavioral change programs in over 60 countries” (emphasis mine). They’ve also reportedly boasted of having influenced elections around the world.
“The information environment has become the new modern battlefield where state and non-state actors employ sophisticated propaganda and disinformation,” reads SCL’s list of services for the defense industry.
This is an organization steeped in the tactics of information warfare. And their clients include governments around the world, and even NATO. (To be fair, not all of their projects are sinister-sounding; they’ve worked on youth engagement, telephone network viability, and food security, as well.)
Cambridge Analytica’s parent company exists to take strong action, through data and messaging, to influence the minds of huge swaths of people. While this isn’t Cambridge, there’s no doubt that the mindset exists in both groups.
Which explains some of the rather disturbing ideas they share in this Channel 4 sting (including using Ukrainian girls to seduce Sri Lankan politicians):
Kogan, Cambridge, and Data Collection
Both Cambridge and Facebook have made a lot of claims. Cambridge knew that they were getting illegally harvested data (or didn’t), Cambridge said they didn’t have any data from Facebook (then said they did), Facebook downplayed the scope of the breach (then didn’t)… it’s a lot to keep up with.
The story here is that Cambridge Analytica used data obtained by Aleksandr Kogan, who had harvested information from 50 million profiles.
Dr. Kogan got that information through a personality test app and a loophole in Facebook’s data-collection rules. It’s important to note, however, that Kogan likely did not violate any rules. Facebook says that Kogan told them the app was for academic purposes, and that selling the data to Cambridge Analytica is a violation of policy.
Kogan has responded by saying that while the app was initially for academic purposes, he later changed the app’s terms and conditions. Facebook says he should have informed the company directly when he did that.
It’s a big, confusing mess.
Who’s at Fault Here?
It’s easy to point fingers in this situation. Kogan sold data he shouldn’t have. Cambridge Analytica bought data they shouldn’t have. Facebook should have been watching for this type of thing. The Trump campaign contracted with a company that has a history of shady practices. And yes, all of those people are at fault. But is a party missing from these recriminations?
Look at it this way.
The data that we generate on social media is an extremely valuable resource. It’s worth billions of dollars. Like oil or gold, people will go to extremes to get it. They’ll buy and sell it. Steal it. Maybe even go to war over it.
But unlike oil or gold, we have control over this particular resource.
Our ambivalence is the root cause of this controversy. We don’t want to face that truth, but we have to consider it.
Yes, Cambridge Analytica used illegally obtained Facebook data. But let’s be honest: they probably didn’t have to. People give away the permissions to their Facebook accounts every day. Remember the people giving the Obama campaign permission to access their Facebook data “consciously or otherwise” back in 2012? That hasn’t changed.
Whether it’s through online quizzes, Facebook apps, web tracking through Like buttons, or just using Facebook to sign into a website, we’re giving away our data all the time. Cambridge Analytica just got it a little faster than they would have otherwise.
It’s been well established that Facebook could be a powerful force, for good or evil, in global politics. Cambridge Analytica had 50 million accounts, but that’s nothing compared to the data that some groups, including Facebook itself, have access to. SCL is not the only group out there offering services like this. Our information is not only valuable—it’s powerful, and it’s dangerous in the wrong hands.
Facebook itself has advertised its power in elections. With over 2 billion users, it has more power than Cambridge Analytica or SCL will ever have. But we’ve remained willfully ignorant of the problem this represents.
There’s possibly no other industry that can benefit more from your data than politics. As long as our Facebook data is valuable—and it’s hard to see that abating—companies like SCL and Cambridge Analytica are going to do whatever they can to get it.
Will we continue to make it easy for them? And will we let them do whatever they want with it?
Assigning blame in this case isn’t easy. The entire system has been headed toward an event like this. It was only a matter of time.
The Future of Data and Elections
The difference between the reactions in 2012 and 2016 is worth thinking about. Was it Snowden’s revelations that taught us to be wary of data collection? The huge data breaches that we’ve seen over the past few years? Or is it because this data collection was done on behalf of Republicans, instead of Democrats?
But when our data starts finding its way into the hands of political campaigns, and when it starts being used to influence the course of national (or even international) politics, many people find that to be a different matter.
Regardless of why we’re talking about it now, it’s something we need to keep talking about. We know that Facebook sells our data. And when they’re selling to advertisers, we generally stay silent.