Recent developments and concerns regarding face-recognition technology
Face-recognition technology has gone from science fiction to ubiquity almost overnight. The biometric technology is used for authentication and identification of individuals by comparing facial features from an image with a stored facial database. The global facial recognition market is expected to generate revenue of $9.6 billion by 2022, growing at a CAGR of 21.3%. However, federal privacy law is yet to catch up with the rapidly evolving technology and there are no standards yet to ensure systems’ accuracy.
Law enforcement use of face recognition
With the help of state driver’s license data, one in two Americans are now in a huge face-recognition database created by U.S. law enforcement agencies.
According to a study released this week by Georgetown Law’s Center on Privacy and Technology, many law-abiding citizens may regularly be scanned without their prior knowledge. Altogether, more than 117 million Americans are subject to face-scanning systems that may be used by police, says The Atlantic. In many states, FBI field offices can search driver’s license photos and police may not need reasonable suspicion to run a search. The Georgetown Law site offers details on face-scanning searches by state in a clickable data visualization.
Facebook is really a “face” book now
Many tech giants, who are commanding a greater role in healthcare, are already using and advancing face recognition technology. Facebook uses face recognition to help users tag friends in photos. Its algorithm is said to be one of the most accurate with 98% accuracy and can even identify you if your face is hidden. It can also identify you quickly, one picture out of 800 million in less than 5 seconds.
Google can help users find photos and videos of themselves, and explains how it uses pattern recognition to make sense of images. Face-recognition tech is advancing artificial intelligence through ‘deep learning.’ Through face recognition, in combination with voice recognition and contextual text analysis, Google and other deep learning systems have access to a lot of personal data.
Among its “Top Ten Strategic Predictions for 2017 and Beyond” for digital disruption, Gartner says that by 2021, 20 percent of all activities an individual engages in will involve at least one of the top-seven digital giants.
Many of us interact with at least one of the digital giants (by market capitalization: Google, Apple, Facebook, Amazon, Baidu, Alibaba and Tencent) in our digital worlds of web search, mobile, social networking, messaging and music streaming. As the physical, financial and healthcare worlds become more digital, many of our activities will be connected and within reach of the digital giants. Note that collectively, the digital giants will have direct and indirect knowledge of what we do as individuals and the fundamental issue will be what they do with the data. – Gartner
Both Google and Facebook are attempting to squash class action lawsuits based on the Biometric Information Privacy Act. Apple’s new facial recognition feature may also spur legal issues. However, Apple’s system is said to be more focused on privacy with stricter limits on how data is collected and stored.
The end of anonymity?
A Russian company, NTechLab, is hoping to make its face recognition technology global, and says anonymity may cease to exist. Its cloud-based API, FindFacePro, is already in demand by over 400 companies, most of them security firms. Its founder, Artem Kukharenko, says American thinking on privacy is insufficiently developed, “Our life will be transparent for anybody. The moral standards will be changed, so it will be a very interesting situation.”
The future of federated patient identification: your face
Innovative health care organizations are using biometrics to support smoother and safer patient access and information sharing. According to Michael Trader, co-founder and president of RightPatient, accurate patient identification is gaining momentum and urgency “due to the rapid digitization of the industry and the concerted push for interoperability and national health information exchange to improve individual and population health.” He says, “The push for increased interoperability could make patient data matching errors and mismatches exponentially more problematic and dangerous and it is widely believed that inadequate patient identification continues to jeopardize patient safety and artificially inflate the cost of care.”
RightPatient’s RemoteID secures remote access to Personal Health Information (PHI) from mHealth apps and patient portals using facial and/or voice verification.
Kairos is another company marketing face recognition for healthcare. It says it is building a platform to quickly add face recognition, and even emotion analysis, into apps and services via its RESTful APIs.
HIPAA requires healthcare organizations to secure remote access to PHI data for patient privacy and to protect against data breaches that can lead to fraud and medical identity theft.
The advent of facial recognition as a unique identifier in a singular or multi-factor environment is a smart answer to the challenge of ensuring a patient receives accurate care throughout the continuum no matter if they are physically present or accessing services from cyberspace. – Michael Trader, RightPatient
Experts say, not one, but multiple layers of security measures need to be coordinated so that if one defense is breached, another is already in place to protect information assets. Multi-factor authentication, for example, combining passwords and biometrics and SMS codes, is said to be the best way to prevent and detect fraud.
Strength of biometrics
Raymond D. Adler, MD, of University of Southern California, presented on biometric technologies at HIMSS16 and outlined the many risks and benefits. Adler says there are many consequences in failing to ID patients accurately, including death. He says the strengths of biometrics include:
- Ability to reliably discriminate among millions of people
- Doesn’t require patient to remember several surrogate identifiers
- Safeguards against intentional misidentification
- Will not change, cannot be impersonated, will remain with the patient over years
Some say face recognition is widely preferred over other biometric technologies – voice recognition, skin texture recognition, iris recognition, and fingerprint scanning – due to its non-contact process and easy deployment. Others say ear scans are more reliable as ears do not change over time.We can expect to see more innovation in the use of biometrics. Click To Tweet
Mobile payments are already being accepted via the Nymi Band with a user’s heartbeat, and Mastercard is also exploring the technology. A fledgling Singapore startup, Aimazing, is close to commercializing a new sound technology solution for mobile payments.
Michael Trader asks, “Is it time to more closely examine implementation of a patient identification solution that leverages biometric facial recognition?”
What are your thoughts?
Latest posts by HealthIsCool (see all)
- Reframing healthcare trends for 2017, see things differently - January 3, 2017
- What ‘skills’ can healthcare develop for Amazon Alexa? - December 15, 2016
- Medicine is about to become data science supported by clinicians - November 22, 2016