It is likely that everyone reading this will have heard of the Cambridge Analytica incident by this point. For those who have not the quick version of the story is that an app designed for Facebook and used most often in conjunction with those personality test seen on the site collected data from approx. 300,000 users which was then used by Cambridge Analytica to create some 50 million profiles on other people that it used for targeted advertising. The legality of this action is currently being addressed with Mark Zuckerberg posting on his own Facebook page that Cambridge Analytica was directed to delete this information but apparently did not. This background information while important is however still background to the more relevant details of the case: how social media and in particular “targeted advertising” may be polarizing us and replacing traditional propaganda.
Now you may be saying to yourself “how is this a big deal? Advertising is hardly new in politics.” And you’d be correct on that account. Advertising is a long used method of getting to voters but it’s only recently that it has taken on such a hidden and manipulative dimension. In times past one could buy ads, go to town halls, hold rallies and send letters all of which where based off of information largely gleaned from prior voting records and polling. Compared to modern methods of information gathering this was akin to pulling teeth and had the inherent issue of selection bias. It did a poor job of illuminating the positions of those who didn’t vote or respond leaving a serious gap in your knowledge base and thus making you vulnerable to sudden turnouts of these groups.
Furthermore, attempts to sling mud at your opponents or play both sides of the fence were on display in the public arena. We saw just such an example of this with Hillary Clinton during her 2016 Presidential election when after pandering to environmentalist about ending coal back peddled on that position when confronted by coal miners which when seen in the public eye is crushing to campaign momentum. Back peddling, double speak, hypocrisy, mudslinging, these are activities that when subject to public scrutiny tend to end poorly for candidates. These very same activities however thrived online during the election and as some would attest are still going strong.
So how does this work? Well it’s complicated but let’s start with the first part, information gathering. The thing you need to know is that when it comes to collecting good and accurate information about humans you need to be clever. For a variety of interesting reasons we don’t really have the time to dive into. You cannot simply go up to someone and ask them what they believe, feel, or think and expect to get the whole story. Asking them what they think about so-so politician triggers a cascade of psychological responses which can taint their output. However, asking them to engage with a silly and seemingly inconsequential personality quiz that tells them which superpower they would have in the Marvel Universe after they say “yes” on the click-wrap agreement which allows you to collect, sell, and access the profiles of them and their friends is a great way to get candid information from them. This, is what Cambridge Analytica did. And if their former CEO, Alexander Nix is to be believe, this gave them unprecedented access to previously unknown political types. Nix went so far as to claim credit for weaponizing the infamous slogan “Crooked Hillary” by trawling the data and finding a group of people receptive to it. Now all this sounds bad and that’s because it is. As I said previously Facebook claims this was illegal and not what they agreed to and that the information was supposed to have been destroyed. But it actually gets worst if you can believe that.
Two words: Dark Ads. Dark Ads were a kind of untraceable, unattributable advertising that could be targeted to certain users that would not appear to those who were not in the target demographic. And when I say “would not appear” I mean you could go to the exact same place as someone in the demographic who saw the ad and it would be invisible to you. The implications behind this become immediately apparent. Using such a system you could, and Cambridge Analytica as well as Russian bots did, create a parallel political reality for people based on targeted advertising that keenly honed in on their political biases revealed through data-mining. As mentioned before some 50 million profiles we created this way keeping in mind that the total voter turn-out was around 136 million in 2016. Now the good news is that such egregious advertising has been stopped by Facebook and Twitter making them no longer invisible but targeted advertising in general has not stopped. And the self-imposed segregation helped by Google and Facebook algorithms which show us more of what we like and less of what we dislike ensures that to a lesser but still concerning degree this happens
Facebook and Twitter may have made those ads visible to everyone but it still requires that one go looking for them if they aren’t in the targeted demographic. And this presents a problem as I can count on one hand the number of people on my Facebook feed who disagree with me politically and I rarely if ever see ads from Republicans or Libertarians. When I do venture out, usually spurred by a story I see from one of my friends, I routinely find myself in an alien landscape where basic facts are different. How exactly is a democracy to function effectively when such divisions exist and are reinforced? If discourse and political debate is the lifeblood of democracy then shouldn’t we, on a national policy level, respond to such naked manipulation and strangulation of discourse? And if this works, which it looks like it does, then you best believe that all the political parties will adopt it as a strategy to the detriment of us who are treated as Guinea pigs in this equation. Often when determining the health of a democracy governmental transparency is a metric employed but perhaps we should extend that metric to the private sector as well given how closely they operate with government in these affairs.
I think almost all would agree with your basic argument that this sort of extreme private political polarization can be and is extremely harmful to democracy, myself included. However, I’m not sure that I understand or agree with your section on information gathering. You seem to suggest that we should be concerned in part because modern methods of information gathering are so superior to older methods of information gathering. This seems to imply that if Cambridge Analytica had received lower quality information about the people they were targeting, its actions would not be so bad. However, I think that the issue with the information gathering aspect was not its effectiveness in ascertaining true views, but that it was done in such an underhanded manner that violated people’s privacy. Even more concerning than the information gathering aspect, however, was the use of that information to micro-target and polarize Americans, as you allude to.
Matthew, I agree that the issue is that such information gathering was underhanded and that is the real issue. I can see how my position on the matter was not clear. I also was using knowledge apparent to me but perhaps not to others that I mistakenly believed adequately addressed my position such as this line here: “after they say “yes” on the click-wrap agreement which allows you to collect, sell, and access the profiles of them and their friends..”
I know I personally find these terms and conditions routinely found in online sites and in other media to be predatory contracts that effectively get people to sign away rights unknowingly because they are ubiquitous (basically every website and business uses them with the same stipulations) and it’s unreasonable to expect someone to read through them much less understand the legal speak found within them ( if I had to retain a lawyer everytime I was asked to click yes to one of these…) and they are very much contracts of adhesion we can’t, for instance as with other contracts, discuss which parts we find unacceptable and come to an agree with the other party.
All this to say I agree with your point that the nature of the information collection and the obvious lack of oversight around such widespread information collection is a problem. A problem I think feeds into other such problems as polarization and when it is utilized in targeted advertising and sorting algorithms which create bubbles around us we don’t even know exist and how surprisingly effective it is. I’ve said before that propaganda is most effective when one doesn’t even realize it’s propaganda and the case of Cambridge Analytica seems to be a case of private entities creating such propaganda for the benefit of government.