Trial of live facial recognition technology to catch criminals is putting people's privacy to the test

Trial of live facial recognition technology to catch criminals is putting people's privacy to the test Sadie Harley Scientific Editor Andrew Zinin Chief Editor WA Police are using live facial recognition technology to help identify and arrest criminals, with significant results. The trial is the first of its kind in Australia. During the three-month trial, nearly 1 million faces were scanned, 209 alerts generated, 114 registered sex offenders identified and 79 arrests made. The trial was considered a success by WA Police, with only eight alerts deemed incorrect. But the use of this technology by police has raised concerns. "When you're widely deploying facial recognition in a live real-life context, you can essentially surveil populations at scale and en masse," says Lauren Perry, responsible technology policy specialist at the Human Technology Institute. "If you don't have very clear thresholds around what it is you're looking for, there is that potential for mass policing. "If I'm walking down the street and I'm getting recognized in the street and arrested because of a parking fine, that's that mass policing sort of state that we really want to avoid." Big Brother The Overt Live Facial Recognition (OLFR) technology compares the faces of everyday people against a predefined alert list containing images of individuals wanted by police. OLFR technology uses biometrics to identify patterns in images of people's faces and determine whether they match a stored image. Despite the images captured being deleted straight away if no match is detected, privacy concerns have been raised. The Office of the Information Commissioner WA (OICWA) said biometric technology like this is considered a higher-risk category of personal information because, once compromised, it is not easy to change. "Transparency and accountability around these risks and how they are being managed is central to building public trust in emerging technologies," the OICWA said in a media release. It also stated that 45% of people surveyed consider facial recognition technology the biggest privacy risk they face today—an increase from 27% in 2023. Policing the police Anna Zenz from the UWA Law School says the privacy laws that apply to this technology have wide-ranging exemptions for law enforcement purposes. "Restrictions that might apply in general to how something like surveillance technology can be used often do not apply when it comes to law enforcement, which is to be expected because they need to be able to do their job," says Zenz. "That means that, from a legal and an oversight point of view, there's not a lot there that provides guardrails, apart from … responsible information sharing and data handling obligations." To help address this gap in regulation, the Human Technology Institute created a model law to provide potential guardrails. "The important thing to point out is that Australia doesn't have any dedicated regulation around the use of facial recognition technology," says Perry. "We developed a model law for facial recognition back in 2022 here at the institute because we realized this is a growing area of concern." "You'd want to make sure that there is a limitation on its use, so you'd want something like a minimum seriousness threshold so that we're only using it to identify and arrest really serious offenders. "You'd want to probably avoid any use of facial analysis, so using facial recognition technology to determine if someone walking down the street is angry or aggressive." Despite the federal government responding positively to the release of the model law, Perry says it hasn't been implemented into legislation. Those guardrails could become increasingly important as researchers examine how constant surveillance can affect people's behavior and well-being. Always watching The University of Technology Sydney recently tested how people reacted to a task while aware they were being monitored. Researchers found the feeling of being watched could heighten anxiety or paranoia. "If we know we're being surveilled by cameras or we have a sense that there might be cameras around us that are watching what we do, the brain sort of subconsciously processes that almost like a primitive risk," says Perry. "We go into that fight-or-flight response." "What's really important for us is to think about this context in the broader scheme of Australia where we're seeing facial recognition technology being rolled out." To limit any nervousness felt by the public, WA Police are making sure people can clearly see when the tech is being used, through signage and posters. Information about deployments will also be made available online before they begin. "These technologies, they're not new," says Perry. "They have been around for a really long time … but they have become introduced into more and more avenues of life." As facial recognition technology becomes more widespread, the challenge will be deciding where the line between public safety and personal privacy should be drawn. Provided by Particle