FB pixel

New York landlords and police uses of facial recognition challenged


A bill has been introduced in New York State Senate to block landlords from deploying biometric facial recognition systems on rental premises, with both state fines and a private right of action for violations, Gothamist reports.

State Senator Brad Hoylman and Assemblywoman Latrice Walker introduced the bill in Albany, following an attempt to deploy a biometric access control system with facial recognition at an apartment building in the Brownsville neighborhood of Brooklyn, part of which Walker represents.

“It was the tipping point for me because facial recognition presents a major privacy concern,” Hoylman told Gothamist. “It gives the landlords the ability to track every entry and exit of tenants and their guests. Fundamentally, tenants should have the right to freely go to and from their homes.”

The State’s Homes and Community Renewal agency must sign-off on the deployment, but has yet to make a decision in the case.

Hoylman also said that while “a more nuanced examination” is necessary before New York follows San Francisco in barring police and other government agencies from using facial recognition, New York should probably consider doing so in the near future.

Georgetown Law’s Center on Privacy and Technology (CPT), meanwhile, has released a report showing that New York police have used altered images and even an image of actor Woody Harrelson to identify suspects with facial recognition.

The report “Garbage In, Garbage Out” details the use of an online image of Harrelson by a detective with the NYPD’s Facial Identification Section (FIS) who noticed that the suspect in a surveillance camera image which had returned no hits looks similar to the celebrity. Harrelson’s image did generate a match, and the individual identified was eventually arrested for petty larceny.

“The NYPD constantly reassesses our existing procedures and in line with that are in the process of reviewing our existent facial recognition protocols,” Detective Denise Moroney told The Verge in a statement. “No one has ever been arrested on the basis of a facial recognition match alone. As with any lead, further investigation is always needed to develop probable cause to arrest. The NYPD has been deliberate and responsible in its use of facial recognition technology.”

CPT also reports that the NYPD has used police sketches with its facial recognition system, which The Verge calls “an unsupported and broadly inaccurate technique.” At least six police departments in the U.S. allow sketches to be used with facial recognition software, according to the report by CPT Senior Associate Clare Garvie. Researchers found correct match rates between 4.1 and 6.7 percent in a 2013 test using an algorithm from Cognitec, which promotes the ability of its technology to match images to sketches, and other tests by NIST and police departments have led to similar conclusions.

Other techniques of dubious value used by police with facial recognition systems according to the report include removal of facial expression, insertion of eyes, mirroring a partial face, combing multiple photographs of similar-looking people, applying a blur effect to an over-exposed or low-quality image, and a “clone stamp tool” which generated an image for part of an occluded face. The NYPD and other agencies are also reported to have used 3D modeling software to virtually rotate faces.

While the results are considered “investigative leads only,” Garvie writes that in the absence of specific guidance, arrests have been made based on little or no other information.

The report concludes with nine recommendations, including that police establish and follow image quality standards, and that they discontinue using “look-alike” or sketched faces.

Related Posts

Article Topics

 |   |   |   | 

Latest Biometrics News


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

The ID16.9 Podcast

Most Read This Week

Featured Company

Biometrics Research

Biometrics White Papers

Biometrics Events

Explaining Biometrics