Cambridge Analytics’s social media campaign is credited by some IT professionals and psychologists for creating the anomalous election last year in which just enough people were surprise voters to tip the electoral college for Trump. The people who describe the psychometric scoring of social media point out the areas in which it violates the ethics of the professionals in IT and psychology, two somewhat ethically challenged fields already.
Berit Anderson and Brett Horvath write for the new long-form online magazine, Scout.
Berit Anderson and Brett Horvath, Scout: The Rise of the Weaponized AI Propaganda Machine
Despite the AI label, the real innovation is applying personality psychology metrics to individual voters as disclosed through their social media advertising profiles and all those silly games that crowd in among the shared and retweeted content.
My recommendation for progressives is not to allow the normalization of this as and extension of the marketing communication of campaigns. We are well through the looking-glass of politicians selecting their voters and winning through multiple unreconciled messages.
By leveraging automated emotional manipulation alongside swarms of bots, Facebook dark posts, A/B testing, and fake news networks, a company called Cambridge Analytica has activated an invisible machine that preys on the personalities of individual voters to create large shifts in public opinion. Many of these technologies have been used individually to some effect before, but together they make up a nearly impenetrable voter manipulation machine that is quickly becoming the new deciding factor in elections around the world.
Most recently, Analytica helped elect U.S. President Donald Trump, secured a win for the Brexit Leave campaign, and led Ted Cruz’s 2016 campaign surge, shepherding him from the back of the GOP primary pack to the front.
The company is owned and controlled by conservative and alt-right interests that are also deeply entwined in the Trump administration. The Mercer family is both a major owner of Cambridge Analytica and one of Trump’s biggest donors. Steve Bannon, in addition to acting as Trump’s Chief Strategist and a member of the White House Security Council, is a Cambridge Analytica board member. Until recently, Analytica’s CTO was the acting CTO at the Republican National Convention.
When winning takes the place of any principles at all, you have a lot of time to figure out how to manipulate voters into your column through nonconsensual means.
It makes one look on Vance Packard’s The Hidden Persuaders with a sense of lost innocence. Indeed, Packard’s whole corpus attains new relevance as we look back to when “America was great”. (From Wikipedia, “Vance Packard”)
1946 How to Pick a Mate – a guide co-authored with the head of the Penn State marriage counseling service
1950 Animal IQ – a popular paperback on animal intelligence
1957 The Hidden Persuaders – on the advertising industry
1959 The Status Seekers – describing American social stratification and behavior
1960 The Waste Makers – criticizes planned obsolescence describing the impact of American productivity, especially on the national character
1962 The Pyramid Climbers – describes the changing impact of American enterprise on managers, the structured lives of corporate executives and the conformity they need to advance in the hierarchy
1964 The Naked Society – on the threats to privacy posed by new technologies such as computerized filing, modern surveillance techniques and methods for influencing human behavior
1968 The Sexual Wilderness – on the sexual revolution of the 1960s and changes in male-female relationships
1972 A Nation of Strangers – about the attrition of communal structure through frequent geographical transfers of corporate executives
1977 The People Shapers – on the use of psychological & biological testing and experimentation to manipulate human behavior
1983 Our Endangered Children – discusses growing up in a changing world, warning that American preoccupation with money, power, status, and sex, ignored the needs of future generations
1989 The Ultra Rich: How Much Is Too Much? – examines the lives of thirty American multimillionaires and their extravagances.
Packard in these books was mostly reporting on recent social science research. Packard would have been 103 this year. A retrospective of his work would be interesting in the current situation as almost all of the players were born within his writing years and not doubt quite a few read him at some point or another.
I can see Steve Bannon reading The Hidden Persuaders and exclaming, “I wanna do this!”
Your case for CA being the Trump secret sauce would be strong if he had flipped more plugged-in districts. Also, Mercer was in the Cruz camp until after he lost.
I can’t pretend to know the mind of Trump voters any more than I can understand why anyone would buy Trump brand clothing or housewares. It’s all stuff designed and manufactured by others and has no intrinsic value or style beyond (or even as equal to) from any store brand.
What does seem obvious to me is that a lot of people in certain locations either didn’t care or voted for change in ’08 and chose to vote for change again in 2016. Maybe the Trump bots beat the Clinton bots, but there was no shortage of the latter.
Like dat.
Bet on it.
AG
With Christie, Huntsman, and Bush in the purple zone, it didn’t leave enough room for Clinton.
Notice…Trump isn’t even the the picture!!!
AG
Trump wasn’t included in the polling during the spring of 2015 which is when I’d guess this graphic was constructed.
Fascinating blast from the recent past, AG. Thanks.
But President Pence is in there. 🙂
Yeah, that would apply as well to David Brock, but Brock is targeting demographics instead of psychological personality types.
The screen of voters into groups is based not on gender, ethnicity, income, and education but these five factors:
inventive/curious vs. consistent/cautious
efficient/organized vs. easy-going/careless
outgoing/energetic vs. solitary/reserved
friendly/compassionate vs. analytical/detached
sensitive/nervous vs. secure/confident
Microtargeting by these factor first gives you target groups like inventive/easy-going/solitary/analytical/secure that you market test ads on, say Facebook, before you deploy them to the specific audience targets, tailored only to the state level geographically. For the electoral college, you only want what works for a state aggregate, regardless of district. So it is the combinatorics of the O-C-E-A-N model that is more important than district- and precinct-level slicing and dicing like the Obama campaign did.
Moreover, because you are tailoring messages to targets, you can run completely contradictory messages within the same geographical space. You don’t need a majority of precinct # zed, alpha, and gamma, but only the aggregate total that wins the entire state. There is no ground game to trace (or the RNC track handled that once the evangelicals were on board).
The other part of it was that a lot of the ads were to depress voter turnout for Clinton, such as by reminding black voters about some of Clinton’s statements trying to straddle the bigots divide between blacks and white working class voters. To deliver this, the target were only certain personality types of black voters who were field tested through earlier ads to act out their responses in certain ways. Not only was sitting out a successful outcome for Trump; so was passing on their misgiving through their personal networks. Likely the ads were not attributed to the Trump campaign. How one could find out from the record is an excellent graduate student (Poli Sci) exercise.
“Personality types” are far more difficult to assess than this formulation suggests. Plus, “personality types” aren’t fixed but vary depending upon time and context. This all sounds more like subliminal messaging, that doesn’t work, then an actual model that can be deployed.
And since it’s not by individuals but the aggregate within a state that the messaging adverts target, it’s no different than what’s long been used by product advertisers. That’s hardly “micro-targeting.” Geographic macro-targeting would be a more correct description.
If one were to target inventive/curious people, you don’t waste time on those in “flyover country.” So, whatever micro-targeting they used to get such people not to vote for Clinton or to vote for Trump didn’t work.
I appreciate that you simply can’t understand how Clinton could actually lose to a bozo like Trump. But Reagan was gaga and he won twice. At a meta-level, voters prefer a tough talking candidate. Especially when the opponent appears to be a wimp or weak and the tough talk appears authentic. That’s how Reagan beat Carer and Mondale and GHWB beat Dukakis. Then GHWB got tagged as the wimp in 1992. Had Gore or Kerry not displayed any wimp qualities, both of them would have beat GWB because his tough talk could be perceived as fake. Hillary is mealy-mouthed and uses a hundred words when twenty would suffice if she had authentic convictions that she’s willing to stand by. Texas and brash NY doesn’t play well on the west coast and worse when the talker is a bozo. Oddly, neither is as big a turn-off in ME and NH as it is in the rest of New England. Trump still can’t believe that he legitimately lost NH, but he did.