US company Clearview AI, which has developed controversial facial recognition technology, has carried out
illegal “
mass surveillance
” in Canada
, according to an investigative report released by the Canadian Privacy Commission.
"
This investigation (...) concludes that the company, based in New York, has violated provincial and federal laws on the protection of personal information
," said the Office of the Commissioner in a
press
release.
Three billion images of faces
"
Clearview's practices represent mass surveillance, which is illegal,
" said Daniel Therrien, the Privacy Commissioner of Canada.
The report notes that the American company has built a database of "
more than three billion images of faces.
Including those of many Canadians and children.
It was accessed by Canadian police forces, as well as private sector organizations, without the consent of individuals, the report laments.
The Royal Canadian Mounted Police recognized in particular that it had used this technology for several months, which can help find strangers, in the context of investigations into cases of sexual exploitation of children on the Internet.
She had stopped using it last summer.
No deletions of photos already collected
The report recommended that the company end this collection, stop offering its device to its customers in Canada and remove the images already collected in order to comply with Canadian law, which Clearview refused.
"
Clearview said its activities have not harmed people
" but can instead help resolve investigations, the report added.
However, the company said it had withdrawn from the Canadian market during the investigation, launched in February 2020. The Canadian regulator had opened it after the broadcast of several reports "
having raised questions and concerns
" on the use of personal information without consent.
International surveys
Last summer, UK and Australian data protection regulators announced the opening of a similar joint investigation.
It focused on the use of the practice of "
web scraping
" used by the company.
This is the automatic extraction of information from a website, without necessarily the agreement of their owners.
Read also: Facial recognition: for better or for worse
On the side of France, a complaint was lodged with the CNIL in July 2020. It was sent by Zoé Vilain, in charge of privacy issues for the French start-up Jumbo Privacy.
The latter offers solutions to better control the use of his personal data.
Zoé Vilain wanted to know "
the extent of the data collected by Clearview AI
".
According to Jumbo Privacy, "
Clearview has been extremely uncooperative in our approach to exercise Zoe's legal rights
."
After four months of discussions, Clearview sent three photos of Zoé Vilain, including one that did not represent her.
“
This shows that they have a problem responding to the
GDPR,
” she explained.
"
And if there are pictures of me, there are potentially pictures of everyone in the European Union,
" she warned.