Just a few years ago, the kind of biometric technology most familiar to everyday consumers was probably fingerprint recognition, thanks to the explosion of fingerprint sensors across a wide range of smartphones. But today, there’s a different modality in the spotlight: facial recognition. And sometimes that light isn’t very flattering.
Much of the credit for the modality’s popularity goes to Apple, which pivoted from its iconic Touch ID fingerprint scanning system to one based on 3D facial recognition in all of this year’s and last year’s iPhone models. Face ID was well-received by consumers, prompting a number of other smartphone brands to deliver their own face-based authentications systems, helping to familiarize a huge swath of consumers with this technology.
At the same time, another major tech company – Amazon – can be credited for bringing facial recognition technology into a different kind of light. What Apple did for face-based authentication, Amazon has done for face-based surveillance.
The ACLU deserves some credit, too. It helped to push Amazon’s biometric technology into the spotlight with its publication of a report in the spring of 2018 in which it revealed that Amazon had sold its computer vision system, Rekognition, to the Orlando Police Department. The report kicked off a firestorm of criticism over the potential for intrusive police surveillance, with even some of Amazon’s own shareholders and employees publicly urging the company to stop selling facial recognition technology to police and government agencies, though Amazon has been steadfast in its insistence that this practice is a force for good.
This was just the beginning of a wider debate. Heightened public awareness about facial recognition-based surveillance has arrived at a time when a growing number of police organizations have been expanding their use of facial recognition in criminal investigations, and exploring its applications in surveillance. And this, in turn, has led to intense scrutiny from politicians and lawmakers. Municipal bans of face-based surveillance systems have started to emerge this year in Silicon Valley, a part of the country known for its zeal for AI-driven technologies; and city officials in other parts of the country are considering bans and restrictions, as well. California’s state Assembly also voted in favor of a ban on the use of facial recognition through police body cameras earlier this year, and this month saw the proposal of a bipartisan piece of legislation aimed at restricting police use of facial recognition nationally; meanwhile, Democratic primary contender Bernie Sanders has made banning the police use of facial recognition an official part of his campaign platform.
Authenticating vs. Identifying
A glance at Chinese authorities’ extensive use of facial recognition in the surveillance of citizens makes it easy to understand this growing concern over facial recognition. But it’s worth noting that this increasingly heated discourse seems to have had little impact, if any, on consumers’ enthusiasm for facial recognition on their smartphones, with Google and Samsung making it a key feature of their most recent devices. This suggests that on some level, there is an intuitive understanding in the public about the difference between biometric identification and biometric authentication. The former is what face-based surveillance is all about: identifying individuals who don’t necessarily want to be identified. Biometric authentication, meanwhile, is the opposite – it’s about verifying the identity of someone who wants to be identified.
The authentication approach still bears some controversy, with civil rights groups like the ACLU have voiced opposition to the government’s use of facial recognition in airports, for example, even though passengers are, by the very nature of the border screening process, volunteering themselves to be identified. But in many other areas, face-based authentication is completely uncontroversial, and even desirable. Major smartphone brands are responding to consumer demand in their launch of new face-based authentication systems for unlocking devices and even authorizing mobile payments; and selfie-based authentication is an increasingly popular means of remote onboarding in areas like financial services, where banks can let new customers open accounts just by uploading images of their faces and official identity documents.
In the enterprise sector, industry pioneers like StoneLock are helping to make facility access easier and faster through the use of facial recognition technology that can enable high-speed processing of large groups of people. Just as smartphone users want to be able to easily access their devices, employees appreciate being able to get to their workspaces with a minimum of hassle, which is what StoneLock’s high-speed facial recognition facilitates.
Leading by Setting Limits
The distinction between face-based authentication and facial identification, then, is an important one. But the growing concerns over biometric surveillance don’t necessarily mean that the latter use of facial recognition needs to be outlawed. Police officials really do seem to view face-based biometric identification as a useful tool, and there are signs that they are willing to work with privacy advocates to find reasonable regulations; in Detroit, for example, city officials recently reached a compromise in banning the police use of facial recognition in live surveillance, but allowing the technology to be used in criminal investigations. Meanwhile, industry leaders like Microsoft and Paravision have taken ethical stands themselves; the former has stated that it refuses to sell facial recognition technology to any government organization that would misuse it in surveillance, and Paravision’s CEO recently revealed in an interview that his team, which is behind a particularly high-scoring solution in NIST testing, “regularly turn down business when we are unconvinced of the responsible use of this technology.” The executive even went so far as to say that it’s unlikely that Paravision will ever operate in China, given the country’s extreme surveillance apparatus.
None of these compromises or ethical stances would be necessary if today’s facial recognition technology weren’t so effective. But with the compelling solutions it offers in both authentication and identification, it’s likely to remain in the spotlight for some time; and with the right leadership, it could really shine from both angles.
November 21, 2019 – by Alex Perala