I just watched the new Netflix documentary The Great Hack. It’s really a deep dive into the Cambridge Analytical scandal and how it impacted both the Brexit campaign and the 2016 Presidential election in the United States.
My overall take is that that movie has some interesting content, but is a bit over-wrought overall. I am extremely interested in the topic but found the movie less captivating than I had hoped.
However, I think the issue it raises — how do we think about privacy and data rights in the modern age — is a critical one for us to work through. And despite the sound bites, mantras, and hot-takes on both sides I think the issue is actually quite difficult and vexing to solve.
However, many of the techniques they used we might consider problematic even being used with the smaller subset of data and potential targets. Frankly, many of the techniques they used are not really new or fundamentally different than previous techniques — they were just far more sophisticated.
For instance, the data allowed them to focus media budgets on relatively small groups of people who could both (a) be induced to change their behavior either by changing their vote or voting when they might have stayed home and (b) were located in an area where such an action might actually change a state-level outcome. Inside Cambridge Analytica, they referred to this group as “The Persuadables.”
Of course, every campaign for a long time has focused its resources on states where the outcome was uncertain (why do you think California gets so little love during Presidential campaigns despite its vast number of Electoral College votes?). Indeed, campaigns have also used a variety of targeting to pick whom to send communications to — and which communications they should send. This has been part of the direct mail playbook for a long time.
So, in many ways, the new techniques aren’t really that different from the old techniques. They are more different in degree than they are in kind. We can do each of them at a larger scale and with much more accuracy than before. I think 3 things have really changed in the environment and politics that may indicate it’s time to treat this change in degree as a change in kind.
(1) There is much more data available on individuals for the purpose of targeting than before. Cambridge Analytica claimed to have over 5,000 data points on every individual American voter. This increase of data is then combined with ongoing feedback (what ads were clicked, which videos watched, etc) and fed through vastly more powerful computing algorithms to make determinations of what to do next. This combination makes for a vastly more effective tool than was available before.
(2) It is now possible to automate the creation of a much wider variety of content — content that is really intended to speak to a very narrow audience. Practical considerations limited this possibility with old technologies (you can’t target TV ads to just one person) — but these new tools allow not just for finding the right people to reach out to, but to tailoring individualized messages to each of them.
(3) There are more parties (some of them directed by foreign governments) who are increasingly willing not only to tailor messages but to intentionally misled or outright lie to the public. This sort of misinformation has proven extremely difficult to prevent (not that I’m giving a lot of credit to the large platforms for trying, but I do believe it is a difficult problem).
When you look at those three changes in combination, it indicated a very different environment rife for abuse and with implications we are only starting to understand and grapple with.
All of which is to say, I think asking if laws were broken is the wrong question. The laws are simply not up to the task of dealing with the media environment that our elections operate in today. We need to be asking the question of how we write laws that can actually bring back some sense of sanity to our democratic process. And that is a much harder question.
Originally published at https://jeffkeltner.com on August 13, 2019.