Computer-vision has made big news in recent weeks in regards to the facial recognition and tagging feature announced by Facebook Inc., as explained in this article published in the Wall Street Journal last month.
Similar to the iPhoto face recognition software launched by Apple Inc. in 2009, Facebook’s new feature suggests whose faces are in a picture. User-added tags enable the new feature to improve its ability over time, as it is “trained” to recognize specific faces more efficiently.
Facebook only allows users to tag photos of people who are listed as their friends, and any user can untag any photos of him or her self. Untagging decreases the chance of Facebook being able to recognize a face in future instances, because it gives the software less information to draw from. Yet there is no guarantee that Facebook servers still aren’t privately collecting and categorizing that data.
This begs the question of what Facebook – or any other entity – may have in mind for the photos and identification data. With more than 600 million users worldwide, Facebook has the capability to create a database of names, pictures and information useful to do things beyond identify friends.
How much privacy are people entitled to when they submit personal information online? Despite Facebook’s assurance of privacy, could the government demand access to such information?
Facial recognition can also be used to identify a suspect in a crowd or a criminal on a security tape, but just as easily misidentify someone, wrongly implicating him or her in a crime. Could these and other potential abuses lead to the eventual eradication of privacy?
The bottom line is this: while Facebook’s new facial recognition isn’t something that people should become hysterical over, it should prompt a discussion of what the possible uses and consequences of that technology are, what is positive and negative about those uses, and what acceptable and non-acceptable uses might be.