Sunday, September 24, 2017

Why Russia Hacked Voter Registration Databases - Micro-targeted Messaging

by Brian T. Lynch, MSW

What if Presidential candidates could identify the exact swing districts, to the precinct, that they need to win state elections? Imagine how concentrated their campaigns efforts and resources would be, including the ad buys and how those ads are tailored to the voters in those districts.

Welcome to the modern political campaign. The ability to do exactly this grows greatly every election cycle. Highly detailed voting information has not only allowed candidates to geographically concentrate their resources, it has given unscrupulous party operatives a map to devise voter suppression strategies, vote tampering schemes and gerrymandered districts that give their party structural advantages.

This much is well known by the savvy readers here, even if it remains under appreciated by the many voter. Less well understood are the new information technology weapons that were employed in the last election.

Into the" big data" world of our modern political campaigns came a whole set of newly developed propaganda technologies that can exploit a campaigns massive knowledge base. Explaining how just one of these new, information technology weapons work, one called micro-targeting, we can see how the dots are connected in the Russia election scandal now unfolding.

Here is a step by step plan to use modern information technologies to micro-target individual voters in swing districts to manipulate their vote.

Step 1. CREATE BIO-PSYCHO-SOCIAL PROFILES ON POTENTIAL VOTERS: All the publicly shared Facebook and Twitter information voluntarily provided by individual users has unwittingly created the most massive database ever imagined. Marketing companies can use this database to target ads to those most likely to buy certain products. But in politics, companies like Cambridge Analytica can use this data to creates highly accurate bio-psycho-social profiles (BPS profile) on millions of American adults, and use that information to manipulate voting behavior. The ability to create these very accurate, highly predictive individual profiles using a meta-analysis techniques is well established. Researchers have estimated that just 150 "likes" on Facebook, along with self-reported biographical information, can produce a BPS profile for individuals that better predicts their behavior than what their own spouse could predict. These profiles can even predict which words or phrases will elicit specific emotional reactions in a person. Of this profiling data, Paul-Olivier Dehaye, a Swiss mathematician, said, "People just don't understand the power of this data and how it can be used against them."

For those who don't know, Cambridge Analytica is a election data analysis company founded by Robert Mercer, an American billionaire with some very radical ideas. The company was lead by Breitbart's Steve Bannon until he joined the Trump administration. Cambridge Analytica was contracted by the Trump campaign to utilize the company's extensive voter profiles to help get Donald Trump elected in 2016.

Information that Facebook or Twitter does not uniformly obtain is the current voter registration status of users or their voting history. This type of information is only kept in state or county voter registration databases spread throughout the country. For a political campaign to get this type of information they would need to hack into many state and county databases, and do it in a way that doesn't easily trace back to the campaign.

Step 2. HACK SPECIFIC VOTER REGISTRATION FILES: Micro-targeting voters is a huge undertaking requiring a massive amount of computing. It also requires connecting an individual's BPS profile with their current voter registration status and voting history. Micro-targeting voters cannot happen without this information. There is currently no national source for voter registration information, but one has been proposed by Donald Trump's Presidential Advisory Commission of Election Integrity.

Not having a national voter registration system is probably a good thing. A hack of a single database would be far more damaging, and is less likely to be detected. Hacking dozens or hundreds of smaller databases increases the odds of getting caught. Hacks of voter registration files should therefore be limited in number, and the hacks must therefore be targeted at just the right swing districts where micro-targeting has the best odds of changing voting behaviors. The people with the best idea of which voting districts to hack are those within the campaign. The same internal polling numbers used to direct ad buys are the same numbers needed to direct voter registration hacks.

One way to assure that hacking activity can't easily be traced back to a candidate's campaign is to covertly employ third parties to conduct the hacks and supply the stolen information to the data analysis companies. If micro-targeting of voters took place during the 2016 elections, as a growing body of evidence suggests, then there has to be a connection between the micro-targeting and the Russian hacks of voter registration files that took place in 22 states. Some information sharing between the campaign and third party hackers would be required to assure that the information to be obtained is useful.

It is still highly speculative, but not unreasonable, to investigate the connections between the Trump campaign and Russian hacking of the voter registration databases. It is also reasonable to investigate whether any of the stolen information ended up in the databases of companies such as Cambridge Analytica.

Step 3. IDENTIFY INDIVIDUALS IN SWING DISTRICT WHOS VOTING BEHAVIOR CAN BE INFLUENCED BY MICRO-TARGETED MESSAGING: Once a voter's registration information and voting history is matched up with his or her BPS profile, it is a relatively straight forward step to distinguish implacable voters from casual or inconsistent voters. BPS profile characteristic can be used to identify a voters political leanings and the issues they might care about. Another characteristic that micro-targeting requires is that the target must be engaged in social media.

Step 4. BOMBARD TARGETED VOTERS ON THEIR SOCIAL MEDIA WITH SPECIALLY DESIGNED MESSAGES: In the final phase of the operation the object is to create an alternative social media landscape for the targeted voter by bombarding them with fake news stories, tweet storms and biased commentary designed to alter their perceptions of the political environment. These messages are tailored to elicit specific emotional reactions in the subjects. The messages are delivered by a virtual army of trolls (Russia has internet troll farms) and automated bots using fake Facebook or Twitter accounts. If the targeted voter ever shared any doubts about Hillary Clinton on social media, for example, the content of their micro-targeted messages might be designed to amplify those doubts and raise new ones. The purpose is to lessen the likelihood of that voter voting for Hillary. If a person ever "liked" a story about building the border wall, targeted messages might contain outrageous immigration stories to heighten fear and loathing toward immigrants, and to strengthen the voters motivation to vote for Donald Trump. By BPS profiling and micro-targeting people, it is the targeted voters who get manipulated, not the voting machines or the voting process itself.


RESULTS: The psychological and emotional impact of targeted propaganda messaging on individual voters will motivate some to go to the polls and vote for a candidate when they might have otherwise stayed home. Or the messaging may dispirit some voters and cause them to stay home when they would have otherwise cast their ballot. Researchers tell us that people manipulated by these technologies generally don't realize they are being manipulated. Because of the massive computing power available to these election data companies, and the unprecedented social media databases, identifying and targeting voters susceptible to targeted propaganda messaging is capable of directing these attacks on many thousands of voters just before an election. Flipping whole election through this process may be possible. Did micro-targeting flip the 2016 Presidential election to Donald Trump's win? No one knows yet, in part because it is so difficult to prove.

It is my belief that the state voter registration hacks were not done to disenfranchise voters at the polls, but to supplement data needed in order to identify and to micro-target low malleable voters with propaganda messaging. I also suspect targeting information was provided to the Russian hackers who broke into the state voter registration files. I don't know if these are crimes. I certainly hope they are, and I hope the Justice Department Probe is pursuing this line of investigation.

No comments:

Post a Comment

Please feel free to comment or make suggestions

Counter