(Photo | Pixabay)
Remember HAL from “2001: A Space Odyssey” which debuted in 1968? HAL-9000 (Heuristically programmed ALgorithmic computer) could recognize who you are, what you’re saying and how you felt about it using facial, voice and sentiment/emotion recognition. HAL could also read lips. It could determine if Dave the astronaut was a threat to HAL’s secret mission, then decide what to do about it. Today, HAL is here. In operation… by nation-states and businesses. Observing, cataloguing, micro-targeting ads, cross-referencing, threat-detecting and spying on the public. Cool and creepy, right? This kind of power in the hands of a benevolent government collecting this data for the good of the people should make us feel safe and secure, but in the hands of a government or company looking out for its own self-interests and ignoring the rule of law… this is what keeps us up at night.
The cool example: Breeze through international flight queues without boarding passes. Your face is scanned as you board, matching it to your passport’s image already stored online. This is already in operation at JetBlue and soon Delta, British Airways, Lufthansa, and American Airlines.
The creepy example: Being tracked because you have an opinion or position which does not align with the powers that be or part of an ethnic group they’d like to marginalize or expatriate.
Sounds kind of dark, but there are also real benefits as well, which is the case for every exponential technology trend. Let’s dive into the challenges and promise.
Less Intrusive Security — Law enforcement already observe and collect info, supposedly under judicial oversight. The public is already used to ICE roadblocks on our southern border, pop-up roadblocks for sobriety checks, truck and auto inspections, DHS airline gauntlets, etc. The promise of this technology is to allow the general public to pass with no disruption while only singling out those which the system predicts/detects as requiring scrutiny. Users of cell phones such as the latest iPhones are already using facial recognition to unlock their phones.
Marketing and Retail — The holy grail of marketing is to micro-target to the individual’s exact needs and wants. Walk into a store and facial recognition identifies who you are and tracks you, and it knows which department to direct you to and which styles you may prefer. It may know if you have a history of buying certain products which it will direct you to and how you feel while you’re shopping. Some stores already use dedicated virtual dressing room kiosks which can show you wearing/using the products and give you the proper size information, patterns and color choices. Combine it with facial recognition, and it will display your own style preferences. Zugara’s WSS for Kiosks virtual dressing platform was demonstrated at the Major League Baseball All-Star Game last year, used by Fanatics, the official MLB retailer.
The positive aspect is providing a tailored experience where you see, touch and buy only the things you’re interested in.
Health — Facial recognition systems, such as FDNA’s DeepGestalt, are diagnosing certain diseases and genetic disorders. A future cybernetic doctor or a selfie on your cell phone may be the first step you take for preventative care, letting you know your body fat percentage, BMI, blood pressure, skin ailments, sleep deprivation, etc. It’s already able to diagnose with high probability certain genetic disorders such as Cornelia de Lange and Angelman syndromes. Other uses in healthcare include more accurate tracking of patients and validating their identity.
Social Media — Facebook is the prime example of using facial recognition technology. Good Use: It can detect your face in photos posted by you, your family and others which is a nice feature for tagging them. Not-So-Good Use: Facebook filed a patent in 2018 for technology to scan photos to build a household profile which can be used to target ads to every member of a family. Facebook could add this “feature” to its existing family targeting program launched in 2017. Talk about ticking off family members…
Privacy — Hands-down, the number one obstacle is the public’s perception of value over cost. The value to society versus the cost of giving up public anonymity. It’s currently at a dull roar but expect this to escalate as the technology spreads and potential abuses occur. How secure is that facial scan? Can hackers grab your digital facial file? What are third parties with access to the data doing with it?
All governments proclaim they are benevolent observers, such as the UK, China, Russia and the USA. It’s not hard to imagine the proclivity to use the tech for nefarious, unlawful reasons, like tracking dissidents, political rivals, or social classes in the ‘wrong’ places, etc. The pushback has begun. In Mid-May 2019, San Francisco became the first major city to ban the use of facial recognition systems by police and other agencies.
What right do we have to control what a retailer does with our digital facial rendering? Can they sell it to the highest bidder, provide it to law enforcement or use it in ads without our consent?
According to the ACLU, Walmart and Target both ran tests in recent years, and both supposedly scrapped its use for now, and Lowes admitted it’s deployed the technology and intends to expand its use for shoplifting and fraud detection.
Accuracy — Chinese law enforcement claim their facial recognition system deployed in public spaces is 98.1 percent accurate. While in the UK, it was found to be up to 98 percent inaccurate. A study by MIT Media Lab and Microsoft Research found that the identification rates for three commonly deployed systems were 91.9 percent for men and 79.4 percent for women, and useless for identifying anyone beyond the binary male/female labels. Accuracy was terrible for gender recognition of women of color, 23.8 percent to 36 percent, and abysmal for lighter-skinned men, 0.0 and 1.6 percent. (source: Wikipedia.org)
Image Quality — Facial recognition works best in controlled environments, such as JetBlue’s queue line where lighting, positioning and range are consistent. Images from even high-end cameras in public places with varying weather and light conditions, obstacles, movement, etc. reduces the image/video quality which leads to lower accuracy. Advances in 3D imagery is dramatically increasing identification rates.
Anti-Facial Recognition Tech — With the clamor over privacy concerns (rightly so), there’s a rapidly evolving anti-facial recognition tech race in progress: glasses with infrared tech which obfuscates the face, hair and makeup tech to confuse the systems, highly-reflective surfaces with distortion patterns and lots of DIY projects proliferating on how to confound the systems.
Once again, this exponential technology can be viewed as either improving our lives or threatening our sense of privacy and security. In the right hands, it makes life easier. I just used the facial recognition on my iPhone to unlock it. On the other hand, government agencies are using it to track us, without our consent. Watch the movies George Orwell’s “1984” … eerily resonant today.
Thousands of retailers are said to be implementing the technology to help prevent fraud and theft, but marketing is the big fish. Usage is far outpacing the ability of policies, laws and corporate ethics to adapt to manage its impact.
I guess I won’t be shopping for lingerie for my wife in stores anymore. Afterwards, I’ll have to explain to her why my Facebook page is plastered with ads featuring lingerie models… using my digitized face and body in the ad from my last visit to Victoria’s Secret. Yikes.
Find Out More
http://bit.ly/Wikipedia-Facial_Recog; http://bit.ly/Facial-Recog-JetBlue; http://bit.ly/Facial-Recog-Diseases; http://bit.ly/Facial-Recog-Retail; http://bit.ly/Facebook-Facial-Recog-Patent; http://bit.ly/Facial-Recog-1984
Preston Callicott is CEO of Five Talent Software, Inc. based in Bend, Oregon. His hope is writing articles such as this one will allow his mind to stop waking him up at 4am with “aha’s” and “oh-my’s” about the massive impact tech has on our collective future.