Before its existence was made public, Clearview AI’s controversial facial-recognition technology sparked such excitement among the Royal Canadian Mounted Police in 2019 that its leaders invited the company to demonstrate it at a Interpol International Conference on Child Exploitation.
Posted at 5:00 a.m.
The US technology, which identifies suspects by comparing their faces to a bank of more than 3 billion images of citizens gleaned without consent from social networks, is described as an “illegal mass surveillance” tool by the Privacy Commissioner of Canada and his provincial counterparts.
Last week, the commissioners issued a joint statement calling for a specific law to regulate the use of facial recognition by the police, justified in large part by the harvesting of images of citizens “in an unreasonable manner” by Clearview AI.
Documents obtained by The pressand thanks to Access to Information Act show that the Royal Canadian Mounted Police (RCMP), when questioned by journalists about its use of Clearview AI in 2020, first wanted to protect this “investigative technique” which had until then gone under the radar. The police force had by then already acquired two licenses from Clearview AI, which it justified by saying that the technology had allowed the US FBI to locate up to 150 child victims of sexual assault by year, whereas this number was previously 50.
Children will continue to be sexually abused and exploited online [si la demande de licence est rejetée]. There will be no one to rescue them because the tool that could have been deployed to save them was not deemed important enough.
Excerpt from an RCMP “request for approval” obtained by The Press thanks to the Access to Information Act.
A few weeks later, an official from the Group for the Fight against Child Exploitation on the Internet (GLEEI) asked a senior company executive to come and give a demonstration of the integration of Clearview into the database of the unity at an Interpol conference scheduled for Lyon or Singapore in October or November 2020.
Investigation
An investigation by the Federal Privacy Commissioner later revealed that the RCMP then authorized the creation of 16 accounts with Clearview without “providing reasonable justification for their purpose or use” or even proceeding to an assessment of the tool’s compliance with the Privacy Act. None of its internal experts in the Technical Investigation Service “were aware of the use of this technology, even though Clearview’s technology was widely used in five different divisions of the RCMP”.
Result: billions of people are found every day, 24 hours a day, in a police identification parade.
Conclusion of the investigation conducted by the Federal Privacy Commissioner
While facial recognition can help ‘solve serious crimes’, commissioners say, it poses ‘too high’ a privacy risk to be used 360-degrees, they said in a joint statement. .
“Even if certain infringements of this right may be justified in specific circumstances, people do not waive their right to privacy”, underline the commissioners. “The current legal framework for the use of facial recognition by police services in Canada is insufficient to address the associated risks to privacy and other fundamental rights. »
With the collaboration of William Leclerc, The Press
A well-founded complaint for The Press
Several details contained in this article come from an access to information request sent on February 10, 2020 to the RCMP. Under the Access to Information Act, the police had 30 days to respond. “The RCMP failed to meet its obligation to respond to the request within the prescribed time frame” and did not provide an explanation as to why it took more than 26 months to do so. ” The complaint [de La Presse] is founded”, concludes the principal director of investigations of the Office of the Information Commissioner of Canada, in a response to a complaint that we officially made… almost a year ago.