Buzz
Facial recognition technology (FRT) is at the center of many debates today. Initially developed in the 1960s, it has only recently become available to both law enforcement and the general public. FRT could potentially replace the need for PIN codes, fingerprints, and even physical keys, but it also brings serious concerns about privacy.
Although it holds great promise in areas like law enforcement, healthcare, and finance, FRT is plagued by issues like racial bias and security flaws. There are even claims suggesting that it could be used to assess one's sexual orientation. With such powerful technology, it’s understandable why many are worried about its misuse.
10. The Racial Bias Dilemma
In the United States, racial inequalities persist in areas like traffic stops, arrests, and stop-and-frisk policies, with African Americans being arrested and surveilled at disproportionately high rates. Unfortunately, FRT tends to underperform with this demographic, leading to a troubling reality: as law enforcement increasingly depends on this biased technology, the consequences are far-reaching.
In 2012, a study involving mug shots in Pinellas County, Florida, uncovered racial bias in facial recognition technology. The system was found to be 5 to 10 percent less accurate at identifying African Americans than Caucasians. One of the companies tested, Cognitec, is already used by law enforcement agencies in Maryland, Pennsylvania, California, and other locations. As Rep. Elijah Cummings, ranking member of the House oversight commission, testified: "If you're black, you're more likely to be affected by this technology, and the technology is more likely to be wrong. That's a hell of a combination."
9. Security Flaws
Apple's iPhone X was marketed as having an unbreakable biometric security system. Yet, just ten days after its release, the Vietnamese security company Bkav revealed that it had bypassed Apple’s facial recognition with a 3D printed mask. The mask, made of a plastic frame, paper eye cutouts, and a silicone nose, cost only about $150.
No one has been able to replicate this feat. Apple denies that it's possible. Wired magazine even employed Hollywood makeup artists to try and duplicate the experiment, but they were unsuccessful. However, simpler methods have proven effective. Researchers at Carnegie Mellon found that accessories like scarves, hats, sunglasses, and even face paint can trick this emerging technology. Cybersecurity expert Stu Sjouwerman of KnowBe4 points out, "Technologies slowly over time learn how to be secure. A pin code or credit card number is something that can be deleted. Your face and fingerprints stay with you forever."
8. Apple’s Major Investment
On December 13, Apple made a major announcement, revealing a $390 million investment in the company responsible for producing the hardware behind the iPhone X’s facial recognition technology. Finisar’s vertical-cavity surface-emitting lasers are crucial for Face ID, Portrait mode, Animoji selfies, the TrueDepth camera, and AirPods' proximity sensors. Apple plans to purchase ten times the amount of these lasers this quarter compared to the total Finisar produced in the last three months.
Apple's huge investment caused a stir in the tech world. Not only did Finisar's stock soar by over 30 percent, but its competitor Lumentum Holdings saw a 10 percent drop in premarket trading. This investment was made from Apple’s $1 billion Advanced Manufacturing Fund, which aims to support U.S. manufacturers and create jobs. Finisar, founded in 1988 and employing 14,000 people, had been struggling after reporting disappointing earnings in the first half of 2017, causing its stock to fall 40 percent.
7. Facial Gaydar
In September 2017, researchers at Stanford University announced they had developed software capable of determining sexual orientation by analyzing facial structure. The algorithm accurately identified homosexual men 81 percent of the time, and when presented with five photos of each man, its success rate increased to 91 percent. For homosexual women, the algorithm had a 71 percent success rate, rising to 83 percent when five photos were used. In comparison, human observers correctly identified male sexuality 61 percent of the time and female sexuality 54 percent of the time.
Researchers Michael Kosinski and Yilun Wang observed that "gay faces tended to be gender atypical." Gay men typically had narrower jaws and longer noses, while lesbians often had larger jaws. Kosinski and Wang theorized that sexuality might be influenced by hormone exposure in the womb, particularly testosterone, which is known to affect facial characteristics.
6. The Future of Medicine
Researchers from Macquarie University in Australia have utilized facial recognition technology (FRT) to assess physiological health. Dr. Ian Stephen, the project leader, explained, "We can determine information about a person’s health simply by analyzing their face, supporting the idea that the face contains valid, perceptible clues to physiological health." Dr. Stephen and his team developed a program to link facial features with BMI, body fat, and blood pressure.
Dr. Stephen’s team tested whether humans could detect health cues in the same way as FRT. When participants were given the option to alter their facial shape, they typically chose to modify their appearance to look thinner, have a lower BMI, and lower blood pressure, aiming to appear healthier. This suggests that, like the computer program, humans process facial features for signs of physiological health, potentially for mate selection purposes.
5. China's Ambition for Global Leadership
China is working toward becoming the global leader in facial recognition technology (FRT). The country is currently building "the world's largest camera surveillance network," with 140 million CCTV cameras already in operation. Over the next three years, the plan is to add 400 million more, many of which will be equipped with FRT.
In China, facial recognition is being used for a variety of tasks. You can use your face to pay for coffee, access attractions, or even withdraw money from ATMs. Some cities also use FRT to publicly shame jaywalkers by displaying their faces and names after they are caught on camera. In December 2017, Shanghai introduced an "intelligent" ticketing system at its subway stations, relying on both voice and facial recognition. This system is intended to reduce congestion at one of the busiest transit hubs in the world, though passengers must first upload their photos to a database for verification.
4. The Facial Recognition Gold Rush
In November 2017, the U.S. Department of Homeland Security hosted a large-scale convention in Menlo Park, pitting private companies against each other to develop facial recognition technology (FRT) for use at U.S. border crossings. The goal was to identify individuals traveling in a car at speeds of 40 kilometers per hour (25 mph) even in light rain. The demand for this technology is so high that they are willing to accept a 70 percent initial error rate.
In January, President Donald Trump signed an executive order to speed up biometric security at border crossings. The emphasis on border security "by any means necessary" signals a potential boom for FRT pioneers. Conference organizers stated, "We want competition. We want them to do better and better." However, not everyone is excited about the collaboration between homeland security and private companies. Harrison Rudolph of Georgetown University’s Center on Privacy and Technology pointed out that many of these systems "aren’t ready for takeoff."
3. Stalker’s Paradise
In March 2016, a Russian company unveiled an app that could identify strangers in a crowd with 70 percent accuracy. The app, FindFace, compares photos to a database of 200 million profile images from a Russian social media site. Developed by Artem Kukharenko and Alexander Kabakov, the app offers the closest match along with ten similar alternatives. Kabakov emphasized that FindFace isn't just for stalking strangers, saying, "You could upload a photo of a movie star you like, or your ex, and find 10 girls who look similar to them and send them messages."
In September 2017, Moscow announced that 5,000 of the city's CCTV cameras would be integrated with FindFace. The technology automatically cross-checks live footage with a database of wanted individuals. After two months of use, Moscow claimed the system helped apprehend six people based on its facial recognition capabilities.
1. Fowl Face
This month, ZhongAn Online, a Chinese company, unveiled a groundbreaking use of facial recognition technology (FRT) in chicken farming. The 'GoGo Chicken' initiative aims to build a database of 23 million birds over the next three years. FRT will enable real-time monitoring of poultry, with a focus on high-end urban consumers who value labels like 'organic' and 'free-range.'
"Food safety is a highly concerning topic," says CEO Chen Wei. He believes that GoGo Chicken aligns with the growing trend of 'farm-based tourism,' which appeals to young city professionals seeking a break from urban life. There’s also a financial advantage to using FRT—by increasing transparency in livestock practices, credit assessment costs can be lowered.
ZhongAn isn't the only company leveraging FRT for animal identification. Recently, Google introduced a feature in Google Photos that uses this technology to identify people's pets, working with both dogs and cats.
