Facial Recognition: Is The Honeymoon Over?
Pressure is mounting for the regulation of facial recognition technology.
PRESSURE to regulate face recognition technology demands the electronic security industry adopt a coherent, thoughtful position based respect for privacy, while continuing to leverage a technology with considerable operational benefits. Whether the right balance can be struck is an open question.
There’s no doubt it’s a tricky time for this powerful biometric analytic. Facial recognition technology is proving powerful, even as an infant in developmental terms. And its potential is vast and lateral, as Samsung’s recent development of a digital clone of Da Vinci’s Mona Lisa in conversation clearly attests. Also important to bear in mind is that the latest facial recognition algorithms are developing exponentially – some have already got around many of the issues for which the tech is being criticised – for instance, a lack of discernment when it comes to skin tones.
In some quarters, negative sentiment around face recognition technology – the view it is a threat to democratic society – threatens to derail use of face recognition technology in key applications. At the same time, large end users, particularly those tasked with protecting public space, are turning to analytics and face recognition technology to efficiently manage operational briefs. It’s not only security managers who appreciate the capability of face recognition technology – law enforcement officers say it saves them time as well.
Regardless of the positive reports from security and law enforcement users, civil lawmakers are not happy. In San Francisco recently, the board of supervisors voted 8-1 to ban the use of face recognition technology by police and other government agencies. The ban prevents city agencies from using facial recognition technology, or information gleaned from CCTV systems that use the technology.
According to Matt Cagle, a lawyer with the A.C.L.U. of Northern California, the technology “provides government with unprecedented power to track people going about their daily lives…that’s incompatible with a healthy democracy.”
In a local case that impacts on all biometric systems, including facial recognition, a Queensland sawmill worker has won an unfair dismissal case against his employer after refusing to register his fingerprints in the company’s biometric access control solution. Though an initial hearing found the company was entitled to install a biometric system, the commission’s full bench later found no valid reason to fire the worker for refusing to provide consent to the company to use his fingerprints and biometric data.
The response of Josh Bornstein, national head of employment law with Maurice Blackburn lawyers, was telling.
“There’s a huge issue more broadly in our society as to whether people’s privacy protections are being maintained, with the rapid pace of technological change,” Bornstein told The Law Report. “We’re seeing employees more closely regulated than ever before — on a 24/7 basis. There’s no doubt regulation is lagging well behind the development of technology…is our personal information – our fingerprint data, the image of our face – property?”
After developing without regulation for decades, face recognition – now powered by deep learning algorithms – may find itself limited in accessing the thing it needs the most – data. The importance of data to useful face recognition was highlighted recently after the ABC published the findings of a Queensland Police Report into the use of face recognition technology during the Commonwealth Games. The ABC’s take on the report had a privacy angle – the further implication was the expense of face recognition had been a waste of money. But QPS’ summation of its experience with face recognition technology using public-facing cameras in static and mobile applications was honest and penetrating. QPS found there simply wasn’t enough data.
“Difficulties were experienced in data ingestion into one of the systems with the testing and availability until the week Operation Sentinel (the Commonwealth Games’ security operation) commenced,” the QPS report found. “…not having the (Commonwealth and state) legislation passed, reduced the database from an anticipated 46 million images to approximately 8 million.”
As well has having very little data to come and go on, the system had so few specific targets it ended up being used for general policing, with only 5 identities of the 268 requested able to be established. Taking this performance criticism to the next level, in the U.S. the House Oversight Committee recently questioned the accuracy and legality of facial recognition tools available to law enforcement agencies and decried the lack of transparency into how the tools are used to monitor public spaces, conduct criminal investigations and identify potential suspects.
Now, in a key development, Republicans and Democrats joined forces across the floor of U.S. Congree to underscore the urgent need to regulate face recognition technology, with calls for a temporary prohibition on government agencies’ using facial recognition technology until clear guidelines are set. Importantly, this is the first time lawmakers in Washington have focused closely on face recognition technology – there’s significant underlying concern about its misuse across the political spectrum.
“No elected officials gave the OK for the states or for the federal government, the FBI, to use [facial recognition],” said the committee’s ranking member, Jim Jordan, R-Ohio. “There should probably be…restrictions. It seems to me it’s time for a timeout.”
Lawmakers on both sides of a political chasm voiced their support for bipartisan action to curb its abuse.
“There’s nothing more American than the freedom of expression and freedom of association,” University of D.C. law professor Andrew Ferguson said recently. “I think what we’ve seen is this kind of technology can chill both of those.”
There’s zero chance regulation of face recognition technology in the U.S. will not flow through to Australia.