Facebook isn’t just a social media platform anymore. It is one of the biggest databases of our lives on the internet. Facebook probably knows more about you than your mother does. It utilizes your data that you voluntarily provide it to empower advertisers to better target you.
So just how much data does Facebook have on you? According to Wired in 2012, before they reduced downloadable data, Facebook had information on you in 57 categories. Some of which include Messages you’ve sent, Profiles you’ve interacted with most, Pokes, Restaurants you’ve eaten at, and more.
With the exhaustive list of data points on your activity on Facebook, it wouldn’t take an average Joe more than a few hours to put together variables to tell an algorithm what to look for to determine a high probability of cheating. And with 9.2 Million Search Results for “facebook probably knows you’re cheating”, it doesn’t seem like we’re too far off.
So I’ve played out the scenario below:
The first determination of probable infidelity, is well, you have to be in a relationship. That’s easy if you’re Facebook official. But what for the people that aren’t Facebook official? You can then scrape images and look for a high density of photo tags, determine proximity, how affectionate the pictures are, length of time in the series of photos. This method isn’t definitive, but can sure get pretty close.
Great, the person is in a relationship, so what’s next?
One of the first signs of online intentional infidelity is when a person begins looking for other options. So you can look to see if the person is spending their time on other people’s profiles, measure the length of time on the profile, time between pictures, likes of the photos, how far back they go back in search, frequency of returning to the profile, and time between “creeping” and messaging, before analyzing the conversation itself.
Since Facebook is already used as a matching method on Tinder and practically every dating app now, it already has the data points it needs to make a good statistical probability of how likely two people are to get along and to what degree. This could mean figuring out mutual friends, likes and interests, activities done on the weekends, political views, etc.
If we start plotting a new chronological timeline when the “creeping” first starts, and plot out activities and data on the timeline, against the chronological timeline of the person and their significant other, we can start to see anomalies and patterns. And since Facebook also has your GPS location, it can detect when two people are in the proximity of each other if they get on Facebook while near each other (which more than likely they will, with 1.3 billion mobile users) and how long they are in proximity… oh and did I mention location (such as overnight stays).
But how do you eliminate wildcards like best friends chatting? Since analyzing the conversation itself would be cheating a little bit, let’s make a bit more interesting. How else could you determine that?
You can measure the conversation against average level of engagement with other Facebook chats, the length of time between responses, type of links included in conversation, frequency of conversation, duration of conversation, and length of average messages.
All of this is to not to say that Facebook actually is monitoring any of this. And there’s probably no way to be 100% certain that someone is having an affair, but there’s a good likelihood that it would be spot on. If Target can learn if you’re pregnant before you do based on what you buy, Facebook can certainly do more with more data points. With machine learning algorithms to learn these patterns, and such a large sample size, a reality TV Show using this data would probably shock even Maury.
So the next time you’re thinking of cheating on your significant other, you should probably use snail mail, or smoke signals. Despite privacy statements that say that companies aren’t using your information, we’ve all signed away our souls on the legal jargon of all the Terms and Conditions that we’ve agreed to. Just remember, everything we do on the internet is always monitored and documented.
This article was originally posted on PixelRocket by Tri Nguyen.