Toronto cops using facial recognition to nab lawbreakers
Facial recognition may have been banned for agencies this month in San Francisco, but in the bustling Canadian city of Toronto, police are using the technology to generate leads in investigations as more and more crimes are caught on video.
Officers have conducted more than 2,500 facial recognition searches since the half-million-dollar system was purchased in March of last year.
HOW DOES IT WORK?
Toronto police use artificial intelligence to compare any photo or video evidence they gather against a mugshot database. That evidence can include anything from government surveillance, to public or private enterprises’ footage captured on security cameras.
This is a controlled use of facial recognition that has obvious benefits for the public. In fact, in the case of the Toronto police, they had 80% accurate matches against their mugshot database; 60% of the time they were able to rely on these matches to further investigations. Though it’s far from perfect, it’s much more accurate than other jurisdictions using this method of identification.
EXTREME USES OF FACIAL RECOGNITION A.I.
Not all police forces are using facial recognition in appropriate ways. The U.K. police are notorious for their Minority Report-esque way of fighting crime before it happens — using software that has a shocking 98% error rate, and that wrongly targets women and people of ethnic minorities.
Beyond the Western world, China has some of the most extreme uses of facial recognition on the planet. An entire city — particularly the Muslim minority population living there — is monitored 24/7 for facial recognition as well as gait recognition (the unique way people walk).
This city had a database breach where all this biometric data leaked out, which illustrates one of the worst aspects of these types of A.I.-driven recognition tools: the fact that once it’s out there, anyone could copy that pattern and abuse it. You can’t change inherent aspects about yourself the way you can easily change a password.
PART OF THE INVESTIGATIVE TOOL SET
Toronto police are describing the way they use facial recognition as “part of the investigative tool set” but not as conclusive evidence on its own — and this is a good approach. It’s not foolproof and involves more error than DNA testing, for example, but because it’s regulated and part of a holistic way of identifying culprits, the risk is limited.
PUBLIC CONSENT NEEDED
Canadian politicians are finally waking up to the importance of controlling any technology that has the potential to infringe on citizens’ privacy rights. Democrat MP Charlie Angus is sounding the alarm on Capitol Hill about the perils of tech that tracks people, saying that as a country we need to discuss guidelines for the legitimate uses of surveillance.
ASK THE RIGHT QUESTIONS
Privacy is all about control and consent. We need to ask the important questions about facial recognition: what level of civilian oversight is provided, how long will the photos and videos be stored, and when investigations are closed, when does the data go away? Being informed and asking the right questions can prevent dangerous uses and abuses of emerging technologies.