My observation is frequently making me think about things that might be not be of much concern to other people as routine matter. It is one of the perks of being very attentive to small things. Or it may be a curse, depending on the situation (it helps in planning).
One thing I notice on a regular basis is that Facebook suggests friends that make you wonder if their facial features really match those of your existing friends. There have been multiple instances where the picture of the person makes it seem like you have known them, although their name does not ring a bell.
The other day I saw a face in my friends’ suggestions that looked too much like one of my old friends who has moved to the UK. He moved there around 20 years ago, when I was still studying at the college level. And after that, I barely had contact with him, and naturally, I wanted to add this person, though on further investigation, I realized that this person was in fact not my friend.
This interaction made me wonder if Facebook was using some sort of face recognition technology for purposes other than photo tagging, like influencing friend recommendations. So, I asked the question from Bing Chat, and this is what it brought up. There are several factors that Facebook uses to suggest friends for you. Some of them are:
- Mutual friends: Facebook may assume that you know someone if you have a large number of friends in common with them.
- Imported contacts: Facebook may use your phone or email contacts to locate potential acquaintances on the network if you have synchronized them.
- Information about your profile: Facebook may utilize details such as your occupation, education, current city, interests, or likes, to pair you with others who have similar information on their accounts.
The above three items do not confirm what I had in mind and have not satisfied my curiosity.
What is Deep Face or facial recognition Technology?
Deep Face is a facial recognition system created by Facebook’s research group that used deep learning to identify human faces in digital images. It was trained on four million photos using a nine-layer neural network with more than 120 million connection weights. On Labeled Faces in the Wild, the Deep Face approach achieved an accuracy of 97.35% ± 0.25%, indicating that it can occasionally outperform humans.
Although it seems that this technology has various benefits for businesses and governments alike, Facebook is constantly challenged by rights groups and privacy advocacy circles. Following multiple challenges, including a lawsuit against Facebook in which a major fine was imposed by the court due to a violation of a privacy law, Meta, formerly known as Facebook, announced on November 2, 2021, that it would be shutting down its facial recognition system and deleting more than 1 billion people’s individual facial recognition templates as a result of this change on its blog.
Privacy Paradox?
Considering that Facebook has closed down its facial recognition activity, why would it be a winner for the privacy of the common people? Governments already have access to ID data in the respective countries of their citizens, which would be a counterargument to the privacy infringement claims.
However, the development of facial recognition technology and its possible uses underscore the significance of continuing conversations about technology, ethics, and privacy.
It can be safely said that there is no proof of any claim to our hypothesis, however, it is a profound possibility that, considering this type of technology exists, it may be used to do this type of stuff. In conclusion, even though there isn’t enough evidence to prove face recognition plays a part in Facebook friend recommendations, it’s a theory worth investigating. In order to make sure that the values of privacy, openness, and ethical use are upheld in our digital interactions, it is imperative that we have constant conversations and close examinations as technology develops.
Another intriguing aspect is how the AI gets its base ideas—is it using stock photos or private data available online? Do let me know what you think in the comments.
References:
Rhapsody 