A facial recognition app that claimed it could track anyone by scanning their Facebook photos has turned out to be a hoax, but it leaves many questions about a future in which similar technology might be reliable.
On March 14th, it was revealed that the facial recognition app Facezam was a mere publicity stunt by a marketing agency based in the United Kingdom.
Facezam said its app had the ability to scan through its database of billions of Facebook photos in a matter of seconds to match photos that the user had taken of strangers on the streets to their Facebook profiles.
“Facezam could be the end of our anonymous societies. Users will be able to identify anyone within a matter of seconds, which means privacy will no longer exist in public society,” said purported app creator Jack Kenyon, while the ruse was still on. Later the marketing form responsible, Zacozo Creative, said it was opposed to such technology.
“Thankfully, face matching apps don’t currently exist in the West. We hope it stays that way,” it posted on the Facezam site.
Facebook, meanwhile, said it would never allow something like Facezam on is platform. It says apps that use automated processes to observe user’s information or collect data are checked to make sure that they adhere to its terms of service before they are allowed to launch.
“People trust us to protect their privacy and keep their information safe. This activity would violate our terms,” a Facebook spokesperson told The Telegraph.
Although Facezam itself was a sham, there are plenty of other tools available which can do similar things and are widely used in places such as airports and other security checkpoints. Even upcoming models of phones are rumoured to use facial recognition to be able to unlock your device.
“Airports in Japan, France, Canada, Australia and elsewhere are increasingly deploying face recognition systems. Most current programs hope to process all passengers at security checkpoints within the next few years,” writes Computerworld columnist Mike Elgan.
In fact, Facebook itself uses facial recognition, you may have noticed the last time you uploaded a picture it try’s to suggest which friends to tag and it’s surprisingly accurate. Popular apps such as Snapchat also use a form of facial recognition for their popular face filters, and there are even elaborate conspiracy theories surrounding the whole thing, if you want to delve down that rabbit hole.
Although Google and Facebook’s facial recognition system are fairly spot on when it comes to a single person or a small group, they as well as many others fall short when a large number of pictures are involved.
Researchers at The University of Washington recently launched a project called The MegaFace Challenge to see if facial recognition software could be deployed at scale. They grabbed nearly 700,000 images from Flickr and tested facial recognition software against it. What they found was previous tests that employed small data sets have worked well, but at such a large scale they failed.
“We need to test facial recognition on a planetary scale to enable practical applications – testing on a larger scale lets you discover the flaws and successes of recognition algorithms,’ said Ira Kemelmacher-Shlizerman, a University of Washington assistant professor of computer science and lead on the project. “We can’t just test it on a very small scale and say it works perfectly.”
Below: Conspiracy Guy falls for Facezam hoax hook, line and sinker: