By B.N. Frank
Warnings about Artificial Intelligence (A.I.) and robots replacing human jobs are nothing new (see 1, 2, 3). Complaints about algorithm inaccuracies and misuse continue to be reported as well (see 1, 2, 3, 4). Nevertheless, Amazon is using algorithms to fire some of its workers.
From Ars Technica:
Amazon is using algorithms with little human intervention to fire Flex workers
Retailer lets bots manage many HR decisions for its Flex delivery program.
Millions of independent contractors are at the whim of a system that Amazon knows is problematic, according to a new report by Bloomberg. While serious early glitches have been worked out, significant issues remain, according to the article. Amazon is reportedly unconcerned about the hiccups and bad press that result so long as sufficient numbers of drivers are available to replace those whose accounts are mistakenly terminated.
“Executives knew this was gonna shit the bed,” a former engineer who designed the system told Bloomberg. “That’s actually how they put it in meetings. The only question was how much poo we wanted there to be.”
The e-commerce giant started Flex in 2015 to handle surges of packages and two-hour deliveries for its Prime Now service. Today, it also coordinates deliveries for Whole Foods grocery orders. (If you’re curious to hear details about how the system functions when it’s working, Ars’ own Sam Machkovech went gonzo a few years ago and worked a shift as a driver.) During the pandemic, the Flex program became an increasingly important part of the retailer’s logistics program, so much so that it offered Flex drivers a bonus to work at least 20 hours in December 2020.
Flex hirings, performance reports, and firings are all handled by software, with minimal intervention by humans. Drivers sign up and upload required documents via a smartphone app, through which they also sign up for shifts, coordinate deliveries, and report problems. It’s also how drivers monitor their ratings, which fall into four broad buckets—Fantastic, Great, Fair, or At Risk. Flex drivers are assessed on a range of variables, including on-time performance, details like whether the package is sufficiently hidden from the street, and a driver’s ability to fulfill customer requests.
Amazon refutes the idea that drivers have been unfairly treated. “We have invested heavily in technology and resources to provide drivers visibility into their standing and eligibility to continue delivering, and investigate all driver appeals,” Amazon spokesperson Kate Kudrna told Bloomberg.
Former Flex drivers interviewed by Bloomberg and driver reports posted on forums across the Internet reveal an opaque system that often leaves them guessing as to why their ratings dropped or why their accounts were terminated—effectively firing them from the job.
A 2019 survey revealed that 82% of Americans believed that A.I. is more harmful than helpful. They seem to be on the right track since there is an actual Artificial Intelligence Hall of Shame with a backlog of submissions that still need to be reviewed. Recently, President Biden committed to introducing even more A.I. into our lives anyway.
Activist Post reports regularly about A.I. and other unsafe technology. For more information visit our archives.
Provide, Protect and Profit from what’s coming! Get a free issue of Counter Markets today.