Face biometrics’ proportionality denied, as Americans become comfortable and Chinese concerns grow
With face biometrics being used at an increasing number of border crossings around the world, some of the systems are being repurposed for unrelated projects, creating huge potential for negative impacts on human rights, according to a new paper from the Canadian Internet Policy and Public Interest Clinic (CIPPIC).
The report, titled ‘Facial Recognition at a Crossroads: Transformation at Our Borders and Beyond,’ suggests that by sharing sensitive biometric data with other government agencies and private companies, border facial recognition systems are introducing new risks. Further, the systems fail the proportionality test, according to the report.
The report authors are also concerned about the impact of demographic disparities in facial recognition, and the accuracy of the systems at scale. Australia is an example in which a government has announced intentions to use travel biometric data available for a wider range of uses, even as project’s like IATA’s OneID seek the expansion of facial recognition internationally, without proper regard for what might be done with the capability, according to the report.
Automated border clearance kiosks are used in more than half a dozen airports in Canada, and the government’s approach to the technology is “characterized by excessive secrecy and few safeguards to prevent repurposing,” the report says.
CIPPIC makes 17 recommendations, from halting the adoption of new facial recognition border control systems to legislating or regulating minimum accuracy and bias thresholds for the technology, along with assessment and reporting requirements.
Report author Tamir Israel tells Vice that the use of facial recognition on smartphones and in social media makes it more socially acceptable, and it can be presented as less invasive because facial images can be collected at a distance.
“Facial recognition is improving in accuracy, but it’s still less accurate than other biometrics and errors are inevitable,” however, he states. As authorities become more confidant in the technology, Israel says they may be less likely to second-guess its errors.
Americans support face biometrics
A new survey of U.S. adults conducted by the Security Industry Association (SIA), meanwhile, shows a majority support the use of facial recognition for a variety of uses, with none more broadly supported than airline applications.
More than two-thirds of Americans (68 percent) say facial recognition can make society safer, 59 percent approve of the technology generally, and 75 percent support its use by airlines. Slightly less support was registered for facial recognition for security at office buildings (70 percent), TSA checks and airport security (69 percent), banks (68 percent), school security (67 percent), and law enforcement (66 percent).
Asked about bias, 70 percent responded that facial recognition is accurate for different races and ethnicities, 54 percent said it can limit human bias, and 47 percent said it could reduce racial injustice and discrimination in law enforcement.
A majority of Americans are also comfortable with having an image in a database used for public safety (57 percent).
“With such significant majority support from Americans for facial recognition demonstrated in the survey, it’s clear the technology is highly valued and recognized for the role it plays in making society safer and improving people’s lives,” comments SIA CEO Don Erickson. “These results indicate that oft-repeated calls by some to ban or strictly limit this technology are clearly out of step with the everyday Americans that would be impacted. Their views must be considered in high-level discussions involving decisionmakers in our communities over how this technology can be used most effectively and responsibly.”
Chinese people’s face data protection concern growing
Incorrectly configured databases are repeatedly spilling facial recognition data from China, even as the technology is rolled out in residential neighborhoods, the South China Morning Post reports.
SCMP recounts the story of a law professor who challenged a neighborhood deployment of facial recognition, and had the system changed to a voluntary alternative to pass cards. The head of the local neighborhood committee told the professor that local police had demanded the facial recognition system be installed.
Face biometric systems are being installed across the country for the stated purpose of combating the COVID-19 outbreak, as well as the traditional security motivations.
At the same time, records that include facial recognition data along with national ID numbers and phone numbers have repeatedly been found for sale online, and at low prices, according to SCMP. Some vendors even offer animation tools to manipulate still images to defeat facial authentication systems.
In the absence of strong regulatory protections for personal data, or significant fines for companies caught treating sensitive data negligently, trust appears to be falling.
“I think there is growing awareness of and pushback against the risks of facial data leaks,” Jeffrey Ding, a researcher with the University of Oxford’s Centre for the Governance of AI told SCMP. He suggests that a data protection law expected to pass soon may address risks from facial biometric data collection. Enforcement, however, he is less optimistic about, due to a lack of capacity among regulators.
Video surveillance market to surpass $33B by 2026
The changing attitudes and persistent concerns are being voiced against the backdrop of a massive and still-expanding video surveillance market. A new report from Fortune Business Insights forecasts the market to grow from $19.12 billion in 2018 to $33.6 billion by 2026, a CAGR of 6.8 percent.
The increasing use of deep learning and video analytics is the main driver of this growth, according to the report, along with smart city projects.
The report also considers the market by component, end-user, and region.