Because of yesterday's Facebook glitch, Instagram and WhatsAPP, some users have discovered how Facebook has discovered the way Facebook has us "packed". What do we mean?
Many images on the website were replaced with their assigned tags from the machine vision systems of the company.
So, if you had browsed through the photos you have uploaded, instead of seeing holiday snapshots or photos of food and friends, you would see a text saying e.g. "image may contain: people smiling, people dancing, wedding and indoor" or simply "image may contain: people smiling, people dancing, wedding and indoor" or simply "image may contain: cat." (= image may contain: cat)
In a nutshell: This is what your life looks like on a computer. This is the way in which the AI Facebook judges you.
The same image tags appear in the Instagram and, in addition to descriptions of scenes and objects, they also suggest who is depicted in a photo based on Facebook facial recognition.
Facebook uses the machine learning to "read" images in this way since at least April 2016 and the project is a big part of the company's accessibility efforts.
What is not clear is whether Facebook also uses this information to promotion of targeted advertisements. There is a lot of data about users' lives that should be protected by Facebook, as it is sensitive information.
In 2017, a developer built a Chrome extension that revealed these labels. "I think a lot of internet users don't realize the amount of information that is usually collected from photos", he said.
Regardless of how this information is used, yesterday's "discovery" sheds some light on the behind-the-scenes of one of the the largest data collection companies in the world.
The improvements in the deep learning the last few years have truly revolutionized the world of machine vision. But how scary could things get in the future?
