Voter targeting becomes voter surveillance
- 12 February, 2016 00:11
Political candidates have always done everything in their power to target voters. But in the current election cycle, with primary election season officially under way, technology is giving them a lot more power than before.
It is at the point where privacy advocates are referring to it as “voter surveillance.”
Bruce Schneier, author, blogger and CTO of Resilient Systems, wrote in his recent book “Data and Goliath” that voter surveillance data can cause “unique harms” to the political process due to, “personalized marketing's capability to discriminate as a way to track voting patterns and better ‘sell’ a candidate or policy position.”
Katharina Kopp, director, privacy and data at the Center for Democracy & Technology, said data collection and analytics can result not only in the “micro-targeting” of some voters, but of others being ignored.
“We should not just ask who is getting what message,” she said. “We need to think about who is not being addressed and what topics are being left out. Who is being left out of the conversation?”
Pam Dixon, executive director of the World Privacy Forum, agreed. “Where are the safeguards against discrimination against poor or minority voters?” she asked. “What kinds of rights do we have as individuals to assert control over how the daily digital exhaust of our lives is used? These are quickly becoming fundamental questions that require answers.”
For the younger voting cohort, this may seem like no big deal. They are already used to retailers knowing enough about them to send them targeted ads. Many of them like it – they would rather see only ads they are interested in, even if it costs them some privacy.
Indeed, the mantra from those in the IT and online security world has been that, thanks to the exploding collection of data on individuals’ locations, activities, social media posts and other online activity, "privacy is dead."
And veteran voters know that candidates have always tended to ignore groups of voters who tend not to vote.
So it may seem like a difference only in degree, not substance, for candidates to use big data analytics to get to know voters well enough to have a better chance of convincing a majority to support them. Even if, to put it more cynically, the effort is to manipulate them more effectively.
They have always tailored their stump speeches, their flyers, their polling questions and their ads to specific interest groups. This is just faster, more efficient and more comprehensive. What is really new?
What is new, and more ominous, according to Evan Selinger, senior fellow with the Future of Privacy Forum and a professor at Rochester Institute of Technology, is what he calls, “an asymmetry of knowledge.
“Average voters have no idea how much information campaigns have compiled on them and how fast a dossier can be updated,” he said. “If they did know, they might object to some of it being taken out of its original context of use, and being put to new use as political fodder.”
That is also one of the major arguments in an article titled “Engineering the public: Big data, surveillance and computational politics” by Zeynep Tufekci, who wrote that while the Internet has enabled much more powerful social movements due to “horizontal communication” that can connect people throughout nations and the world, those same digital technologies, “have also given rise to a data-analytic environment that favors the powerful, data-rich incumbents.”
He said big data analytics can, “foster more effective – and less transparent – ‘engineering of consent.’”
“Computational politics,” he wrote, amounts to, “significant information asymmetry – those holding the data know a lot about individuals while people don’t know what the data practitioners know about them.”
If people did know, there is evidence that Selinger is correct – they would object. The title of a 2012 paper published by the University of Pennsylvania’s Annenberg School for Communications summed it up rather bluntly: “Americans Roundly Reject Tailored Political Advertising.”
According to a survey conducted by the authors, 86 percent of adults said they, “do not want political campaigns to tailor advertisements to their interests.” The respondents said if they found out they were being targeted in that way, it would make them less likely to vote for the candidate doing it. By contrast, only 46 percent objected to targeted discounts from retailers.
A somewhat smaller, but still large, majority – 70 percent – said they would object to a campaign using Facebook to send ads to the friends of a person who “likes” the candidate’s Facebook page.
But, of course, if they don’t know about it, they are less likely to object. Kopp and others say that is another problem with the collection of personal data by political campaigns – it is largely invisible.
“Part of the problem is that we don’t know enough about these practices,” she said. “The whole process is opaque.”
Tufekci made the same point. He said modern data collection is dramatically different from that of a generation ago, when tracking what magazines a voter reads or the kind of car he or she drives, “required complicated, roundabout inferences about their meaning … and allowed only broad profiling.”
Big data, he wrote, allows for much more individualized profiling, and since it, “can be collected in an invisible, latent manner and delivered individually,” it means a candidate’s organization doesn’t have to ask questions of a voter to know a lot about that voter.
That, he wrote, turns political communication into a more personalized, and less public, transaction that is also vulnerable to “subterfuge and opacity.”
Dixon agreed. “It would astound people to know how our daily actions and even our thoughts about candidates can be predicted from data science – or at least, that is what the sales pitch is,” she said. “It is arguable that data analytics are more predictive than traditional voter polling.”
Selinger said that lack of awareness can make voters more vulnerable. Indeed, the Iowa secretary of state criticized the Republican winner, Ted Cruz, for sending out a mailer just days before the vote, aimed at driving voters to the caucuses by giving them poor grades based on their voting history and accusing them of a “voter violation.”
“We fail to appreciate how inexpensive our personal information has become to political campaigns, so we can drop our guard a little when campaigners communicate with us in the personalized ways – ways we otherwise only expect from people whom we're intimately connected to and associate with,” Selinger said.
So far, all of this opaque data collection is also essentially unregulated. “We’re sort of at the beginning of this,” Kopp said. “There are academics studying it, but there is no real regulation. And there aren’t a whole lot of people working on it.”
There is also the potential security problem. Colin J. Bennett, in an article titled “Trends in Voter Surveillance in Western Societies,” wrote that sensitive voter data, “can be put in the hands of multiple volunteers and campaign workers, who may have no privacy or security training. In a world where data breaches are commonplace and daily occurrences, the decentralization of voter intelligence data could be a disaster waiting to happen.”
But, it appears there is little likelihood of anything changing during this election cycle. Kopp notes the obvious – politicians aren’t all that interested in passing a law that would restrict the effectiveness of campaigning.
But she and others say at a minimum, voters ought to know specifics about it. “There should be greater transparency about what data is collected, how long it is retained, and who it is shared with,” Selinger said.
And Dixon said there should be “stringent security requirements” placed on political campaigns, including, “access controls and data retention and deletion policies.”