Amazon launches low-key campaign to derail Portland face-scanning law
Independent news reporting indicates that Amazon is spending money to defeat proposed Portland, Oregon legislation that would ban all use of biometric facial recognition systems — even those privately owned — except for a short list of proscribed roles.
Two national technology publications monitoring the evolving situation in Portland say Amazon is soft-pedaling its campaign to stop any ban.
Redtail, an online publication of Pandion Media LLC, has reported that an Amazon executive sits on the board of the think tank Information Technology and Innovation Foundation (ITIF), which has written an anti-ban opinion article in the local paper, The Oregonian.
The foundation lists Shannon Kellogg, Amazon’s vice president of public policy, as one of its members. People in similar positions at face-scan players Apple Corp. and Microsoft Corp. also have memberships with the foundation.
OneZero, part of the Medium journalism site, has reported that Amazon spent $12,000 last year lobbying against the proposed law. The publication said that this is the first time Amazon has lobbied Portland.
Leaders of the iconoclastic city has been considering restrictions of facial recognition which some citizens feel invades their privacy. And there is evidence that at least poorly crafted algorithms more frequently misidentify people with darker skin, potentially exposing them to undue attention from police and wrongful discrimination by businesses.
The legislation, according to OneZero, would allow the technology to be used to “unlock a phone, tag someone in social media, or to obscure faces in images and video.” It began as a ban on use by city government agencies.
Executives unhappy with a possible ban say face recognition systems can make their places of business safer and reward loyal customers. Advocates in government say they want to use the systems to prevent and solve crimes.
Were it adopted, the proposal would make illegal a face-scanning system that’s been deployed by local outlets of Jacksons Food Store, a regional convenience market chain. Entry to the outlets is possible only if a facial recognition system cannot match someone with the stored image of a criminal who had previously victimized a Jacksons store.
Amazon’s product, Rekognition, has been making a name for itself, but not always positively. An American Civil Liberties Union in 2018 carried out a study of the Rekognition’s capability using photos of members of Congress. Twenty-eight members were erroneously matched with people who had been arrested for a crime, according to the study. Amazon dismissed that study claiming that the default confidence setting used by the ACLU was inappropriate to the use case.