Psychographic data (personalities and traits beyond simple demographics) has long been known and used before Facebook ever existed, it's just much easier now that so many people are feeding data to a centralized system with easy access.
Worldview Standard has a podcast called RawData where ep 1 of season 1 in 2015 talked about how a few likes on Facebook would let you know someone better than their own family, based on research from 2013: http://worldview.stanford.edu/raw-data/episode-1-uploaded
None of this is new. It was only either ignored or accepted before and has finally reached critical mass due to the amount of controversy and conspiracy today, along with the generally expected fatigue of social media and the now evident effects it makes on most people's lives.
Interestingly, according to the NYT Cruz campaign staffers did not think it worked:
"But Cambridge’s psychographic models proved unreliable in the Cruz presidential campaign, according to Rick Tyler, a former Cruz aide, and another consultant involved in the campaign. In one early test, more than half the Oklahoma voters whom Cambridge had identified as Cruz supporters actually favored other candidates. The campaign stopped using Cambridge’s data entirely after the South Carolina primary.
“When they were hired, from the outset it didn’t strike me that they had a wide breadth of experience in the American political landscape,” Mr. Tyler said."
This is in some sense an extension of the Myers-Briggs-type personality typing, which in its early days also grew out of data analysis.
People use the Myers-Briggs typing at work because it helps you work with (read persuade) people.
So you can make an argument this is at least 100 years old.
As for why we've reached critical mass - it seems likely the ability to influence democratic elections, and the efforts by enemies of the Western world to use it to undermine democracy, are getting people to notice.
No, the history of personality typing is a much more interesting story than that, and depends very specifically on data.
Myers/Briggs/Jung identified 4-5 personality "axes" on which people vary. How did they do that? They basically gave a bunch of people surveys with hundreds of questions and did PCA on the data. They found that 4-5 dimensions explained a lot of the variance. That was a new finding based on data, at the same time that the field of statistics was developing. And it gave insight into how people behave. It's important enough that we frequently use it as a heuristic in workplaces today.
In modern times, we can do the same on much larger data sets. Given Facebook "like" data for 50 million people, you can do dimensionality reduction on the data and extract personality "types". There's no question that this gives you information about people. The question is how well it can be weaponized - that's the debate around CA now.
> As for why we've reached critical mass - it seems likely the ability to influence democratic elections, and the efforts by enemies of the Western world to use it to undermine democracy, are getting people to notice.
American politicians of both stripes have realized the private sector now wields comparable propaganda power to that of the government, hence the post election "Russia" propaganda activity burst and full court press on the story by friendly newspapers to rein everyone in, before they are neutered.
Yes, I think that's right - Western societies need to think very carefully about the role of data, the power of private corporations, and the tension between profit-seeking and societal goals. Government regulation is the way we align societal goals with individual goals like profit seeking.
But the Russian state does spend billions on their secret services, and they consider America their "Main Enemy". Billions buys you something. I'd be shocked if there weren't more shoes to drop about Russia. But for now, we should be focused on domestic disinformation.
CA's CEO is seen presenting this clearly in 2016, specifically how well it helped the Cruz campaign: https://www.youtube.com/watch?v=n8Dd5aVXLCc
Worldview Standard has a podcast called RawData where ep 1 of season 1 in 2015 talked about how a few likes on Facebook would let you know someone better than their own family, based on research from 2013: http://worldview.stanford.edu/raw-data/episode-1-uploaded
Obama's campaign used heavy data analytics for both runs, explained as early as 2012: https://www.technologyreview.com/s/509026/how-obamas-team-us...
None of this is new. It was only either ignored or accepted before and has finally reached critical mass due to the amount of controversy and conspiracy today, along with the generally expected fatigue of social media and the now evident effects it makes on most people's lives.