Why I Care About Privacy

Common responses

I care about privacy because I want to take back (most) control of the digital information we ingest and, ultimately, how we think, purchase, and make decisions. While trying to understand my own interest in the topic, I went through the same arguments that I've heard elsewhere:

All my data is out there anyways

Translation: "it's too late and I'm lazy". Just because your current and past behavior has been mined doesn't mean you need to throw away your rights to your future data as well. There are easy, free things you can do to greatly mitigate the risks of releasing your private data in the future.

I have nothing to hide

Translation: "I don't know what I should be hiding". It's not just that you "like sports". It's that the speed at which you type and how many times you use autocorrect when you message your friend after your team loses the game indicates an agitated mental state before bed, which increases your likelihood to click on an ad for ZZZMelatonin in your morning feed . The extent to which your data is mined and flipped into actionable, seamless recommendations is only going to increase.

Long term risk prevention is the real hard consequence to motivate, and it's the most important. Every time we accept cookies or allow apps to use location services or load a site without an ad blocker, we are choosing convenience over security. Security against what? Data leaks are the easiest to point to and the easiest to dismiss. There's a reason car insurance is required by law. Data leaks and car crashes are distant and potentially catastrophic. What concerns me more is the subtle, continuous nudge that has the ability to change how we think.

Decisions

We are faced with millions of decisions and our decisions shape our reality. Where to live, what to do for work. who to spend time with. But also smaller decisions like where to eat. "Making a decision" implies picking one out of many options. Now, we are well past a world where we are aware of all the possible options for a given decision, so we come up with heuristics to narrow the search:

• Any restaurant over 4.5
• Best General Tso for takeout
• Taco truck near me

Often, we have so many options in a completely different domain than we're used to that instead of our own heuristics, we rely on algorithms.

Where heuristics were built by us using our desires to filter out choices in the real world, algorithms represent a much more expressive space of functions. At best, algorithms try to infer the user's preference and present options that match. At worst, algorithms influence a user's preference based on other heuristics completely out of the user's control, all while presenting favorable to the user themselves.

Algorithms

Search, ranking, social media, TV, news all rely on algorithms to whittle a deluge of options down to a set that resonates with the heuristics of their users. These options are increasingly diverse in almost every dimension, so firms need an algorithm that’s innately adaptable while optimizing for a single optimizable metric. That metric only sometimes has to do with the user’s preference. For platforms that serve ads, the metrics are: length of time spent on their platform and Click Through Rate. Both of these metrics bubble up into Revenue:

$\frac{\#\text{clicks}}{\#\text{views}}$ is known as the Click Through Rate (CTR). In order to increase CTR, the ad company (Google, Facebook) needs to show the ads to people who are more likely to click them.

At first that seemed pretty simple. Based on posts you've liked, shared with friends, or saved, I can build up an understanding of your preferences and target you with ads accordingly. As the predictive power of that internal platform data was exhausted, and continuous revenue growth was demanded, advertisers started pulling from external metrics. Location, search history, browser history, where you live, income, demographics, content of private messages and emails, facial expression when you see a provocative tweet. Each morsel of data gets thrown into the algorithm and nudges the CTR ever so slightly higher. And since it is performing at scale, 0.1% increase in CTR could result in $100m of additional revenue. With deep pockets and a long tail to chase, ad providers have been on a hunt for new and unique ways to mine personal behavior data.

The last part of that equation, multiplying by number of views, concerns the owner of the advertising surface (FB, IG). The longer they can keep you on their platform, the more they can show you ads and collect your in-app data. So, with another algorithm, these ad surfaces optimize for "time spent on app", aka "attention". Algorithms answer questions like: What post should I show next? What liked by ___ username should I display? Which comments should I prioritize in order to extend the expected length of the session. It turns out that this problem also benefits tremendously from granular personal behavior data.

With two tremendous sources of demand, the personal behavioral data industry is booming, served by data vendors whose sole job is to construct unified digital fingerprints of everyone online. As that dataset coalesces, more demand floods in. Whether you’re selling ads (Google, Facebook) or selling products (Amazon, Walmart) or making trading decisions (banks, hedge funds), knowing exactly how a person is feeling and interacting with the digital world is extremely lucrative.

Personalization to Exploitation

The initial purpose of the algorithms was to replace heuristics as options for a decision (what to watch, where to eat, what to read, what to pay attention to) become vast and diverse. These algorithms stray from “estimating what the user wants” towards “guiding what the user sees” as they explicitly optimize for profits, using user preference as an auxiliary goal. The more personal data they can mine and purchase about you, the better that guide becomes at clearing a path between what you want and what is best for the company.

If we make more and more of our decisions based on choices filtered by these algorithms, we have decided to be shepherded into a reality we did not consciously choose. While significant legal work has to be done to fix the systemic issue, tools are available to remove yourself from the wholesale of your personal behavioral data and the inferences companies make from them. These steps may not be convenient and you may not see the effects of them immediately, but they take you on a path to reclaiming your autonomy online rather than incrementally ceding more of it.