Implications of Facial Recognition and the #TenYearChallenge
#Challenges have become part of social media culture. We have seen the Tide Pod Challenge cause kids to eat laundry detergent, and the Bird Box challenge has led to teenagers crashing cars on the freeway as they emulate the mega successful Netflix movie by attempting to drive cars blindfolded. The latest challenge to hit social media is called the Ten Year Challenge. In this dare, people post a current picture with one from ten years ago to see how hard they have aged. Sure there is a bit of narcissism in the craze, but many are beginning to question if there are more conspiracy driven reasons for wanting age based images of the population: data mining to train facial recognition algorithms on age progression and age recognition.
Facial recognition is quickly becoming mainstream and even many cell phones have the technology pre-installed as a way to unlock your device. Sure, it's handier than a pin code or swipe pattern but there are some privacy concerns being raised about giving away such a massive amount of facial data. Even the mainstream Wired chimed in on the potential implications of the Ten Year Challenge.
In the world of data, context is important and miners try to eliminate 'noise' - data which doesn't further define the set of information. So when you post a picture of your dog as a profile pic, or a building you visited, 'noise' is created in the data set which makes 'you' harder to define because it lacks the context a computer needs to comprehend it. By labeling your photo '2018' or '2008' a computer and data miner is better able to see what the data means and assign value or relevancy to what is being mined.
Now, for the sake of argument, it's true that most of us have already uploaded plenty of photos to our social media accounts meaning it's data that is already available but there is no real timeline for the images. Facebook, Instagram (Facebook owned), and Twitter don't really know how old we are in our photos. By posting and clearly labeling one as current and one as ten years old we are placing valuable data right in their hands. The questions then becomes how can they use it?
Facial recognition software can have positive uses: police in the Indian city of New Delhi used the technology to search for 45,000 children throughout the city. In just four days they were able to identify nearly 3,000 missing children and reunite them with their families. Giving a computer extensive data to work with could help it better identify aging curves and predict the growth and appearance of children as they age, thus making it easier to find more kids and missing persons.
The potential implications of facial recognition technology in the healthcare field are enormous. One healthcare app called Face2Gene is able to detect certain phenotypes through specific facial and non-facial features a patient has which it uses to detect some genetic disorders and syndromes.
Facial recognition has allowed the gaming industry to make strides: applications from Luxand FaceCrop SDK are able to use your face to create 3D avatars for games or build AR (augmented reality) worlds based on your real-time, real-life face transformations. These same applications have enhanced the intelligence of cameras which can now recognize faces and tag them for later reference as well as automate post-processing tasks such as skin tone adjustment and red-eye removal for optimal photography.
Another use for the technology is in the blossoming world of home automation. As more people work to build 'smart homes' advancements could help to create key-less, face recognizing security systems and cameras.
The downsides to being recognized on camera center around the arguments of invasion of privacy and presumption of guilt. Facial recognition could mean we all receive labels and classifications. Are you a potential trouble maker? Predisposed to a certain disease? What if a certain feature was identified that suggested you might become a murderer or rapist? Could that be used as a reason to send you to jail before having ever committing a crime? If it sounds like a sci-fi story, consider this:
In 2016, Amazon introduced real-time facial recognition services called Rekognition (sounds like they were watching us through our webcams) and quickly began selling those services to law enforcement and government agencies, including the police departments of Orlando and Washington County, Oregon. Sure, police could use that information for good and track down violent criminals - but they could also use the data to identify potential criminals. In essence guilty until proven innocent. This could include protesters, or those who are just viewed as a nuisance. So there is a potential to abuse the power bestowed by facial recognition.
The American Civil Liberties Union recognized the downsides to Amazons service and asked the company to stop selling this service. Amazon’s shareholders and employees also requested a halt to the service, sending a letter to CEO Jeff Bezos citing concerns for the company’s valuation and reputation. Bezos replied by saying that he would continue doing business with big government agencies, including the US Department of Defense."If big tech companies are going to turn their back on the DoD, this country is going to be in trouble." Still trust Amazon or believe they have anything but profit in mind?
Though the medical applications of facial recognition could help pinpoint disease before its too late, that could also mean the insurance industry will quickly jump on board in hopes of mitigating risks. Those who are identified as having underlying medical conditions could be charged higher rates or denied coverage completely. Of course, those identified as potential criminals would no doubt seek out cosmetic surgery to hide from the camera that might be law enforcement looking for a reason to take one into custody on the presumption of guilt.
Facebook has been previously caught red handed when it comes to mishandling of profile data: the company was hacked in 2018 and data for as many as 70 million users - which included search history and location information - was released. 15 million users had sensitive data like their name, email address, and phone number exposed while another 14 million had additional profile info like gender, religion, location, device info, locations you’ve been tagged in, and Pages you’ve liked released to hackers. Another 1 million users had access to their accounts stolen, but never used. Facebook said that hackers did not post anything on users profiles as far as they know.
For what it's worth, Facebook denies having any part in the #10YearChallenge. A spokesperson said "Facebook did not start this trend, and the meme uses photos that already exist on Facebook. Facebook gains nothing from this meme (besides reminding us of the questionable fashion trends of 2009). As a reminder, Facebook users can choose to turn facial recognition on or off at any time.”
As far as we are concerned, we don't think there's much that can be done to prevent the implementation of facial recognition software. Yes, they are plenty of beneficial uses that will hopefully one day make the world a better place. But until the bugs are worked out we will continue our own personal social media blackout.