Hired by Machine, Fired by Machine: a Heartless Amazon

Stephen Normandin spent almost four years racing around Phoenix delivering packages as a contract driver for Amazon.com Inc. Then one day, he received an automated email. The algorithms tracking him had decided he wasn’t doing his job properly.  The 63-year-old Army veteran was stunned. He’d been fired by a machine.

Normandin says Amazon punished him for things beyond his control that prevented him from completing his deliveries, such as locked apartment complexes. He said he took the termination hard and, priding himself on a strong work ethic, recalled that during his military career he helped cook for 250,000 Vietnamese refugees at Fort Chaffee in Arkansas.

“I’m an old-school kind of guy, and I give every job 110%,” he said. “This really upset me because we're talking about my reputation. They say I didn’t do the job when I know damn well I did.”  Normandin’s experience is a twist on the decades-old prediction that robots will replace workers. At Amazon, machines are often the boss—hiring, rating and firing millions of people with little or no human oversight.

Amazon became the world’s largest online retailer in part by outsourcing its sprawling operations to algorithms—sets of computer instructions designed to solve specific problems. For years, the company has used algorithms to manage the millions of third-party merchants on its online marketplace, drawing complaints that sellers have been booted off after being falsely accused of selling counterfeit goods and jacking up prices.

Increasingly, the company is ceding its human-resources operation to machines as well, using software not only to manage workers in its warehouses but to oversee contract drivers, independent delivery companies and even the performance of its office workers. People familiar with the strategy say Chief Executive Officer Jeff Bezos believes machines make decisions more quickly and accurately than people, reducing costs and giving Amazon a competitive advantage.

Amazon started its gig-style Flex delivery service in 2015, and the army of contract drivers quickly became a critical part of the company’s delivery machine. Typically, Flex drivers handle packages that haven’t been loaded on an Amazon van before the driver leaves. Rather than making the customer wait, Flex drivers ensure the packages are delivered the same day. They also handle a large number of same-day grocery deliveries from Amazon’s Whole Foods Market chain. Flex drivers helped keep Amazon humming during the pandemic and were only too happy to earn about $25 an hour shuttling packages after their Uber and Lyft gigs dried up.

But the moment they sign on, Flex drivers discover algorithms are monitoring their every move. Did they get to the delivery station when they said they would? Did they complete their route in the prescribed window? Did they leave a package in full view of porch pirates instead of hidden behind a planter as requested? Amazon algorithms scan the gusher of incoming data for performance patterns and decide which drivers get more routes and which are deactivated. Human feedback is rare. Drivers occasionally receive automated emails, but mostly they’re left to obsess about their ratings, which include four categories: Fantastic, Great, Fair or At Risk.

Amazon knew delegating work to machines would lead to mistakes and damaging headlines, former managers said, but decided it was cheaper to trust the algorithms than pay people to investigate mistaken firings so long as the drivers could be replaced easily.

Inside Amazon, the Flex program is considered a great success, said a former engineer who helped design the system. “Executives knew this was gonna shit the bed,” this person said. “That’s actually how they put it in meetings. The only question was how much poo we wanted there to be.”

In a statement, Amazon spokeswoman Kate Kudrna called drivers’ claims of poor treatment and unfair termination anecdotal and said they don’t represent the experience of the vast majority of Flex drivers. “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,” she said. 

As independent contractors, Flex drivers have little recourse when they believe they’ve been deactivated unfairly. There’s no paid administrative leave during an appeal. Drivers can pay $200 to take their dispute to arbitration, but few do, seeing it as a waste of time and money.

When drivers do challenge poor ratings, they can’t tell if they’re communicating with real people. Responses often include just a first name or no name at all, and the replies typically apply to a variety of situations rather than a specific problem. When human managers get involved, they typically conduct a hasty review—if they do one at all—because they must meet their own performance standards. A former employee at a driver support call center said dozens of part-time seasonal workers with little training were assigned to oversee issues for millions of drivers.  “Amazon doesn’t care,” a former Amazon employee said. “They know most people will get their packages and the 2 or 3 percent who don’t will get something eventually.”

Legislators at the state and federal level have studied the matter but have been slow to enact rules to prevent  harm. In December, Senator Chris Coons, Democrat of Delaware, introduced the Algorithmic Fairness Act. It would require the Federal Trade Commission to create rules that ensure algorithms are being used equitably and that those affected by their decisions are informed and have the opportunity to reverse mistakes. So far his proposal has gone nowhere.