The technology giant Amazon is providing police agencies and private intelligence contractors in the United States with facial recognition and video tracking algorithms to conduct mass surveillance on the population.
Amazon’s artificial intelligence software, named Rekognition, transforms CCTV, police body camera and other government video sources into an omnipresent eye, constantly collecting data and analyzing it in real-time using databases of tens of millions of images of the population. It provides police agencies with the power to track the movements and activities of targeted individuals and identify individuals in public places such as at demonstrations.
The program was brought to public attention by a report published by the American Civil Liberties Union (ACLU) on Tuesday. It was accompanied by an open letter from the ACLU and other civil liberties groups to Amazon’s CEO and the world’s richest man, Jeff Bezos, calling on Amazon to cease providing the service to government agencies.
Amazon Web Solutions (AWS) unveiled Rekognition in November 2016, marketing it as a useful tool for government agencies to “inspect photos for objects or people of interest or concern.” One year later, Amazon announced the addition of a video monitoring feature, which “detects persons” captured on video, “even when the camera is in motion and, for each person, returns a bounding box” around their face, identifying “face attributes.”
The AWS website states: “For security and public safety applications, this helps identify persons of interest against a collection of millions of faces in real-time, enabling timely and accurate crime prevention.” The program can also identify “vehicles based on license plate numbers from images taken by street cameras.”
The company also announced a new feature, “Crowd-Mode Face Detection,” which can identify 100 individual faces in a single image. It states that the program can “capture demographics” and “analyze sentiments” based on facial analysis of “group photos, crowded events and public places.” The blog post includes an accompanying image of hundreds of people whose faces have been automatically detected.
Both the Washington County Sheriff’s Office in Oregon, and the city of Orlando, Florida, are publicly listed as current Rekognition customers on the Amazon Web Solutions (AWS) website.
The operations in Orlando were described at a recent developer conference in Seoul, South Korea by the director of Rekognition, Ranju Das. “City of Orlando is a launch partner of ours,” Das said. “They have cameras all over the city. The authorized cameras are then streaming the data…. We analyze the video in real time [and] search against the collection of faces they have.”
Emails obtained by the ACLU under a freedom of information request submitted in January show that Washington County sheriffs introduced the software by February, 2017, within three months of its unveiling, and immediately combined it with a database of 300,000 police photos. The county also created a mobile app for its deputies to submit images for verification while on duty.
Another email obtained by the ACLU, dated July 10, 2017, shows a request from the Oregon Terrorism Information Threat Assessment Network (TITAN) Fusion Center to process images through the sheriff office’s Rekognition software. Fusion centers are under the control of the federal Department of Homeland Security and coordinate the sharing of information on “counter-terrorism” between the FBI and state, local government and private contractors.
The publicly-listed users of the Rekognition software also include ARMED Data Fusion Systems, a private intelligence contractor dedicated to “the development and integration of cutting-edge technology to combat acts of political violence, terrorism, organized criminal activities, and insider threats.” The contractor’s website lists a number of services, including responding to “mass public demonstrations and civil unrest,” and identifying “radicalized individuals who pose an imminent threat.”
During a speech to a four-day AWS re:Invent conference in November 2016, the chief data scientist for Motorola Solutions, Dan Law, promoted the company’s new software developed in conjunction with Amazon’s Rekognition. A video demonstration of the software shows police body-cam footage being live-streamed and processed to identify every person in the officer’s line of sight in real time. When a “missing” suspect is identified, it immediately calls all nearby officers to the scene.
The integration of artificial intelligence technology, controlled by giant technology corporations run by billionaire oligarchs, into the repressive apparatus of the capitalist state, poses the gravest threats to the democratic rights of the working class.
It is not hard to imagine to what uses this technology will be put as mass opposition continues to grow in the working class in the US and internationally to the policies of austerity and war offered by the financial elite. Individuals participating in anti-war demonstrations are to be automatically identified and targeted. Immigration agents are to be empowered with the ability to immediately identify immigrants targeted for deportation.
Workers in factories are to be subjected to an ever-greater level of corporate surveillance, monitoring for “disobedience” and repression. In a speech delivered on January 23, 2018, AWS’s senior product manager, Venkatesh Bagaria, gloated that Rekognition should be adopted by “factories who want to really understand what is happening on their floor.” To remain “operationally efficient,” he said, “they could use recognition video to look for particular activities or certain people. If a worker is not supposed to be in a particular area and he shows up, you could find him with facial recognition.”
Amazon’s Rekognition program is only the latest demonstration of the integration of the giant technology corporations, including Google, Facebook and Amazon, into the military-police-intelligence apparatus. Amazon’s AWS has hosted a top-security data center for the 17 American intelligence agencies that make up the “intelligence community” since 2013, when it won a $600 million tender for the contract.
Last April, a letter to the CEO of Google, signed by more than 3,000 Google employees, demanded the ending of the company’s “Project Maven,” which involves the use of artificial intelligence to analyze objects in drone footage for the Pentagon, and could easily be used to identify targets for drone assassinations across the Middle East and Africa.
At the same time, Google and Facebook are assisting the US government to suppress political opposition to these policies. Since April 2017, both have announced measures, in the name of fighting “fake news,” aimed at reducing access to socialist, left-wing and anti-war publications.
The author also recommends:
Google, drone murder and the military-intelligence-censorship complex
[23 May 2018]