Don’t Just Delete Facebook, Scramble Your Data First
If you’re savvy with code, you can employ a script that repeatedly alters your Facebook posts with nonsense, making it more difficult for the social media site to collect user data.
Kevin Matthew, a former systems administrator who owns a small web development company, shared a script he created that replaces existing Facebook posts with randomly-generated nonsense. With a little coding know-how, you could use this script to repeatedly mangle all your Facebook posts over a period of several months, to make the bulk of Facebook’s data on you virtually unusable (though it doesn’t do anything for the data that’s already been scraped by third-parties, like the kind Cambridge Analytica allegedly gained access to).
Matthew said his script is simply a proof of concept, because actually doing this may be in violation of Facebook’s terms of service (so use at your own risk).
Why purge your Facebook data when you can just delete your account?
This is the first question I want to address. For anyone who has worked as a systems administrator, DevOps or developer for highly available enterprise environments you should know that data replication, geographic redundancy, backups, offsite backups and data retention policies are standard best practices that one would expect to find in these types of scenarios.
For Facebook, which accounts for about 40% of all referral traffic on the internet, you can only begin to imagine the replication, scaling, redundancy and other strategies that are employed across multiple geographically redundant data centers that Facebook operates.
What does this all mean?
This means that even by conservative assumptions, your data never really disappears permanently if you deactivate and delete your Facebook account. If you’re lucky enough to live in the European Union, then you might have better chances with the right to be forgotten. In North America, I don’t see any reason to assume that your data is actually permanently deleted when you delete your account. It might even be safe to assume that this data is held and transparently linked to any new accounts you might open in the future, either by connecting phone numbers associated with accounts or by algorithmic statistical analysis.
How do you purge your facebook data before deleting your account?
Since we will never know the extent of data retention policies in the depths of Facebook’s proprietary engineering strategies, we can take the paranoid approach.
No strategy outlined here or anywhere else can purport to be 100% effective, however, anyone with a good Systems Administration background can appreciate that the paranoid approach is the safest.
With that in mind, what we want to do is “poison” or obfuscate all our data such as timeline posts, likes, comments, and any other information that we have submitted to Facebook’s systems. What this means is that all these metrics that we have submitted to facebook (willingly) are part of what facebook uses to generate all the analytics and scary algorithmic calculations that determine things like who we voted for, sexual orientation and any number of things that we may have not willingly disclosed.
So we go back in time and obfuscate the data. All of it.
What I would propose, with the CasperJS script, is to automate the editing of all status updates on Facebook with random or pseudo-random data (i.e. “spun” text that is readable but randomly generated). These status updates would need to be edited and re-saved multiple times over the course of an undisclosed amount of time.
We want to edit every post and replace it with random data (for the sake of argument) 5 times, over the course of 3 months. So if you had 1,000 posts in your facebook timeline, you would need to make 5,000 edits over the course of 90 days. Then we would want to have the data “sit” on Facebook’s systems to ensure that the re-saved data has poisoned even the historical backups as best as we possibly can.
This sounds extreme, but we want our replaced and obfuscated data to reach all of the Facebook endpoints in order to properly anonymize the data, ultimately taking back what we did not willingly give. That is to say, whatever Facebook has computed against all our actions to determine who we are as people so that they can better advertise to us.
How to use the script?
Head to the original article here.