The New York Police Department has drawn the ire of privacy advocates after misleading the public about its use of Clearview AI’s controversial facial recognition system. While the NYPD has acknowledged that some of its officers had trialed the platform in the past, it denied any formal arrangement or consistent use of the technology.
However, new documents obtained through freedom of information requests show that those statements at best misrepresented the truth and at worst were little more than outright lies. NYPD officers have in fact performed more than 5,100 searches using the Clearview platform, and in many cases used the technology in an active investigation to lead to an arrest.
The emails also show that the NYPD and Clearview had an extremely friendly relationship that dates all the way back to October of 2018, when NYPD deputy inspector Chris Flanagan was introduced to Clearview CEO Hoan Ton-That. NYPD signed a formal vendor contract in December of 2018, and actively encouraged its officers to use the platform as much as possible at that time. That initial trial ended in March of 2019, but the department continued to communicate with Clearview and open new trial accounts for nearly a year afterward, with the most recent ones created in February of 2020.
The specific way in which the NYPD used the technology was similarly deceptive, insofar is the department seems to have used Clearview to circumvent established policies that govern its use of facial recognition. At the NYPD, the use of facial recognition is supposed to be restricted to a single team, and that team is only supposed to perform searches with pre-approved software, and are only supposed to reference a supervised image repository.
With Clearview, the NYPD handed out access privileges to an unregulated database to many officers that were not part of the facial recognition team. In some cases, the department even allowed officers to run the Clearview app on their personal devices instead of those issued by the department, which are subject to more oversight.
The NYPD’s actions may have also violated New York’s status as a sanctuary city. As a sanctuary city, New York is not supposed to coordinate with federal agencies on immigration enforcement, but officers promoted Clearview to (and may have shared information with) members of Department of Homeland Security (DHS) task forces. Immigration and Customs Enforcement is a part of DHS, and became an official Clearview customer back in August.
The freedom of information requests were filed by the Legal Aid Society and journalist Rachel Richards. Clearview is already facing lawsuits in multiple jurisdictions for its invasive data collection practices.
The NYPD, meanwhile, is not the first law enforcement agency to try to cover up its use of facial recognition. The Los Angeles Police Department was recently forced to acknowledge that its use of the technology was far more widespread than it had previously admitted, though it did ban the use of third-party facial recognition systems after learning that some officers had used Clearview on a trial basis.
Source: MIT Technology Review
April 14, 2021 – by Eric Weiss