Last week, it came up in class that Facebook now uses facial
recognition in helping the user to tag pictures. Ever since the first time I
witnessed it happen, I have thought it was unbelievable, but have never been
too concerned about it. I just did a little research, and apparently I am
unusual for not being more worried that Facebook has this ability. For the past
couple of years, there has basically been a battle between Germany and Facebook
about its use of facial recognition. About a month ago, the conflict was
resolved. This resolution caused Facebook to delete facial recognition data for
its European users, both Irish and German.
What exactly is it,
though, that these countries are worried about that wasn’t brought to my
attention? Data protection laws in Europe state that users must give
specific consent for their images to be used in the way of facial recognition.
The basic concept of facial recognition is that analytical software to compile
archives of human faces. Facebook, however, uses what is considered an “opt-in”
system, which assumes users will automatically want to use facial recognition.
How exactly does
facial recognition work and what are its current challenges? Facial
recognition faces current challenges in scalability of databases. There are so
many obvious factors that we can think of that make it difficult for this
technology to work. Changes in lighting, makeup, and the environment are a few
that come to mind. (Facebook somehow didn’t have trouble picking out my face
with or without my big sunglasses in my Spring Break pictures, though!) Facial
recognition systems first detect face positions and poses and extract them from
the image. The difficulties discussed above create bottlenecks in the
computation. The social media area of facial recognition opens up a new door
because so many images are put into the image database, which makes it much
easier to confuse similar faces.
Sources:
I agree with you that many of us are turning a blind eye to the rapid growth of a "big brother" type phenomenon with social media. Not only is Facebook bringing to light the facial recognition systems that we thought were only Hollywood TV show propaganda, but also the notion that publicly available software has been approved by the government (exactly what European governments fought against). What information is stored about us in the vast information source of the world wide web? How easily can our whereabouts be traced by not only government entities, but also normal citizens by simply analyzing Facebook and following where we are tagged in photos. Another similar issue is with apps such as FourSquare or other "check in" mobile apps that allow users to publicly post their locations. How much information are we freely giving away through social "games" and interactions? Not necessarily all of this information is harmful or self incriminating, but how far will things go until privacy is no longer a possibility?
ReplyDeleteOn a completely different argument however, you could use the facial recognition system that Facebook has incorporated to increase user privacy to a whole new level. What is stopping current users from posting pictures not of themselves and hiding behind a false identity? Could the facial recognition system confidently identify the person in a photo and require approval before the photo can be posted publicly? This would certainly reduce the occurance of undesired pictures to be posted. This often happens with celebrities and paparazzi. Could you eventually control the distribution of any photograph taken of you through this facial recognition? It is interesting to see this technology could lead to two very different possibilities of personal control to privacy.