By Kaitlin Durbin Jul 21, 2011, 3:06 GMT
Washington - Does automatic digital face-recognition pose a new threat to internet privacy?
That is the question being raised by the Washington-based Electronic Privacy Information Center (EPIC), which has filed a privacy complaint against Facebook with the Federal Trade Commission (FTC), the chief privacy policy and enforcement agency in the US.
EPIC said that Facebook's new automatic facial-recognition feature forms the base for digital face-searching that eventually could be used internet-wide.
The public interest research centre has already notched one success on its belt concerning internet privacy issues; its complaint to the FTC last year forced Google to shut down its attempted social networking site Buzz. Google had been automatically using the personal email addresses from Gmail users to build a networking web for Buzz.
Now, Marc Rotenberg and his EPIC team are going after almighty Facebook, charging that facial recognition on the world's most popular social networking site has opened a whole new Pandora's box.
'People agreed to share their photographs with friends. Facebook decided to take those photos and build a biometric database that would allow facial-recognition of Facebook users,' Rotenberg told the German Press Agency dpa.
The European Union, which regulates privacy in Europe, has also filed a complaint against the feature.
Facebook's automatic facial-recognition feature has astonishing capabilities; it uses biometric summary data to take measurements of each individual face in an uploaded photo. Each face is assigned a numeric value based on the measurements, and the value is used to compare similar facial features across multiple photos.
Faces whose numeric values match are grouped together and Facebook suggests the name of the pictured individual. If the uploader agrees that the face goes with the name, then a simple click of confirmation will officially link the user's name and Facebook profile to their face in the photo.
The technology was developed in 1988 by Professors Lawrence Sirovich and Mike Kirby.
Television series like CSI have led the public to believe that the use of facial-recognition technology is common in today's law enforcement, but actual enforcement bureaus like the FBI have declared the technology unreliable for actual use.
Despite EPIC's complaint, Facebook dismisses any suggestion that the new feature treads on internet privacy.
Andrew Noyes, manager of public policy at Facebook, said that with more than 100 million tags added to photos every day, tag suggestions were simply meant to make photo tagging easier.
Assistance, rather than invasion, was the goal, Noyes said. 'No action is taken on a (user's) behalf and all suggestions can be ignored,' he told dpa.
Noyes said that Facebook subscribers can disable the feature in their Privacy Settings if they do not want their name to be suggested to other friends.
But in reality, disabling the feature does not restrict users from tagging friends manually, nor does it stop the facial-recognition grouping process. It simply stops the feature from suggesting that the particular user's face belongs to their name.
Prior to the feature's debut, users had to manually tag each friend in every photo. Facebook users do not control which of their friends tag them, but they do have the ability to untag themselves from a photo.
Untagging does not remove the photo from a friend's photo album, but it does remove the photo from the user's personal account.
Noyes said the new feature, which offers the same untagging abilities as the manual version, has generated 'almost no user complaints.'
For Rotenberg and EPIC, though, convenience does not constitute right.
'Facebook has basically inserted itself in the middle of the tagging and untagging camps and said to people 'Let's do a lot more tagging. We're going to help you tag more people, we're going to show you who to tag, we're going to give you the names for tagging.' And that does seem to upset the balance,' Rotenberg said.
'This is really Facebook trying to encourage people to make their friends' images readable by Facebook,' Rotenberg said.
Rotenberg fears that the facial-recognition tool could eventually make all Facebook photos, regardless of privacy settings, 'name-searchable,' meaning Internet users-friends, family, hackers, employers-could type a name into a search engine such as Google Images to pull up every photo of the person on record, including the collection on Facebook.
'This is not just making it more convenient to tag your friends in your photos,' Rotenberg said. 'Images (could) become searchable by name because the anonymous face is now essentially machine readable because a Facebook ID has been assigned to (it).'
If Rotenberg is right, this new feature could have even more dire effects on users and job seekers who have already suffered because of their Facebook pages.
According to a 2009 study by CareerBuilder, 45 per cent of US employers look at a candidate's Facebook page before hiring. And 8 per cent of companies with 1,000 or more employees have dismissed a worker based on information found on Facebook, a 2009 study by Proofpoint, an Internet security firm, said.
Keith Crosley, director of market development for Proofpoint, said that the current percentage of firings due to Facebook was likely higher.
'I expect that (percentage) to go up in our 2011 survey as many more organizations now have social media policies and, as such, one can expect more enforcement actions, both in terms of discipline as well as terminations,' Crosley said.
Your Talkback on this Story