LAST December, Ed Bridges was mingling with the crowds of Christmas shoppers on the streets of Cardiff, UK, when the police snapped a picture of him. He has been trying to get them to delete it ever since.
Bridges hasn’t been convicted of a crime, nor is he suspected of committing one. He is simply one of a vast number of people who have been quietly added to face-recognition databases without their consent, and most often, without their knowledge.
For years, critics have warned that the technology is an unparalleled invasion of privacy, but the rise of face recognition seems unstoppable. Police forces across the world have launched face-recognition programmes, setting up cameras to scan crowds at football matches, festivals, protests and on busy streets in a bid to identify criminal suspects.
Advertisement
The tech giants are also in on the game. Facebook relies on face recognition to . Snapchat uses it to . The latest iPhone ditched fingerprint scanners in favour of using face recognition to unlock devices. Amazon’s , among other things, that it can spot faces from a library of suspects for law enforcement.
It is surprising just how far this tech reaches. US airline JetBlue is trialling it to . Visitors to Madison Square Garden in New York are . Jaywalkers in Shenzhen, China, and having their mugshot taken by cameras placed at pedestrian crossings. It seems nobody can escape.
“Visitors to Madison Square Garden in New York are scanned to see if they might pose a problem”
Yet around the world, face databases are running into problems. India’s supreme court is due to rule on whether its national ID scheme Aadhaar . The system – which includes mandatory enrolment in a face-recognition programme – contains details on , and has suffered huge leaks of sensitive data.
In the US, an found that it had built up a database of more than , including half the US adult population, without proper oversight. Suspected criminals made up less than 10 per cent of the library.
Meanwhile China is using face-recognition technology to monitor and discriminate . As well as scanning people’s faces before they enter markets or buy fuel, the system alerts authorities if targeted individuals , effectively building virtual checkpoints to hem in Uighur Muslims.
Such discrimination aside, you might feel reassured about having your face stored in a database if it helps solve crime. However, despite its popularity with law enforcement, face recognition is often inaccurate (see “graphic”).
Figures released by the South Wales Police show that when the technology was , 12 people were flagged by the cameras as matching the police database of faces, but only two of the matches were correct.
Similarly, the American Civil Liberties Union (ACLU) recently demonstrated how Amazon’s Rekognition erroneously identified members of Congress in a database of unrelated mugshots. It is now calling for a moratorium on the use of the technology by law enforcement.
“There’s a worry the tech will spread, and start to be used before we’ve had the debate about whether it should be used at all,” says Neema Singh Guliani of the ACLU. “This is very imperfect technology, going into institutions that are themselves imperfect. That combination makes problems, which should not be swept under the rug.”
Critics warn that innocent people are being put at risk by software that wrongly identifies them as criminals. “In very simple terms, tech is outpacing legislation,” says Paul Wiles, the UK’s biometrics commissioner, who is charged with overseeing the government’s use and retention of DNA and fingerprints.
In the UK, these biometrics were restricted by a law passed in 2012, but there are no such restrictions on technology such as face recognition, voice recognition and iris scanning, which have taken off since then. “There’s been a significant development of biometrics as part of a bigger development of capacity to store and use data,” says Wiles.
After being photographed shopping in Cardiff, Bridges was snapped again in March, when he joined a protest against an arms fair being held in the city. “At lunchtime this face-recognition van suddenly appeared across the road from the main group of protesters,” he says. “I felt it was done to intimidate us, so we would not use our right to peacefully protest.”
“There’s a very real concern face recognition will contribute to a surveillance structure”
“Protesters, activists, will think twice if they know when they speak out about government abuse they’re going to be recognised,” says Guliani. “There’s a very real concern face recognition will contribute to a surveillance structure, where people don’t feel like they can walk around with anonymity and privacy.”
Bridges is now taking legal action to challenge the use of face-recognition technology by South Wales Police. Assisting him is Megan Goulding, a lawyer at the human rights group Liberty. “Part of the reason we’re challenging the use of face recognition at all is we think it’s pretty impossible to protect yourself,” she says. “Because of the indiscriminate nature of the technology, it can happen without your knowledge or consent.”
Liberty is pushing for a judicial review, a legal mechanism that allows individuals to challenge the actions of public bodies. Goulding hopes that a legal case would rule that the way face recognition is being used breaches human rights and data protections laws, leading to a halt in its use.
Face the future
If the South Wales Police are found to have acted unlawfully, it would prevent other police forces using the technology in the same way. Goulding expects a ruling early next year, but Wiles is cautious about relying on a single decision. “I think the danger is the government responds on an ad-hoc basis, face recognition today, voice recognition tomorrow, and so on. What they should be doing is setting out a strategy for the police in general.”
Similar conversations are happening elsewhere. In July, regulation on the development and use of face-recognition technology. “I think we need trials that protect privacy but allow development,” says Wiles. “These should be peer reviewed, published, all the usual things you expect from a medical trial.”
Such trials would demonstrate the reliability and limitations of the software, such as whether it is less accurate when matching people from minority groups. Only then can a sensible public debate about the technology take place, says Wiles.
Wiles points to the approach of the Scottish government. In July it published a for the use and retention of biometric data. A commissioner can then draw up codes of practice for each application, so that different rules can apply for matching the faces of people who have been arrested to mugshots compared with automatically scanning faces in crowds. “It’s a clever solution to what seems an insoluble problem,” says Wiles.
However, this is a compromise too far for Goulding. “In our view we don’t think it’s possible to balance the existence of this technology. The impact on rights is too grave.”
This article appeared in print under the headline “Invasion of the face snatchers”
