This article is more than 1 year old
Uber drivers game Uber's system like Uber games the entire planet
App cabbies push back against controlling black-box computers
Uber drivers are resisting Uber's algorithmic management to raise their wages and to push back against uncompromising computer control.
As described in a study presented earlier this year at the Theorizing the Web Conference in New York and the 4th International Workshop on the Sharing Economy, in Lund, Sweden, researchers Mareike Möhlmann and Ola Henfridsson, from Warwick Business School, and Lior Zalmanson, from New York University, gathered data from 19 transcribed interviews and from more than 1,000 posts made to the UberPeople.net website, a forum used by many Uber drivers.
They found that Uber drivers have organized mass "switch-offs" to trick Uber's software into activating surge pricing – a response to the apparent lack of available drivers. They also found that Uber drivers have tried to make UberPOOL pay better.
Uber encourages drivers to accept UberPOOL ride requests by setting its commission at 10 per cent for the first UberPOOL passenger, instead of the 30 per cent commission applied to UberX rides.
So to collect a greater portion of the fare, Uber drivers have tried picking up a single UberPOOL passenger and then logging off the system to avoid picking up additional passengers for what's supposed to be a shared ride.
Uber's software now attempts to fight this by automatically adding a second pickup to UberPOOL rides, according to a forum quote excerpted in the paper.
An Uber spokesperson insists misbehavior is rare and is deterred through policy and code. "This behavior is neither widespread nor permissible on the Uber platform, and we have technical safeguards in place to help prevent it from happening," the spokesperson said in an emailed statement.
To some extent, people can be expected to look for ways around technological obstacles to self-interest. Efforts to bend or break rules have existed throughout history.
But Möhlmann, Henfridsson, and Zalmanson see Uber's attempt to manage people through code as particularly problematic because the company's lack of transparency has led drivers to believe the system is rigged against them.
"The drivers experienced a loss of autonomy mainly in cases where they felt the system was not transparent and when they felt they were not being treated 'fairly' or were not receiving full and relevant information," the paper states. "In such cases, these unknowns hampered their ability to plan ahead and maximize their value capture."
Ride-sharing platforms need to share information
Management decisions by black box algorithms generate resentment, the researchers suggest. In theory, they say, algorithmic management has the potential to increase system transparency by being built on a known set of rules and procedures. But Uber's strategy, they contend, has been to withhold information.
"Our study shows that Uber's strategy is perceived very negatively by drivers, and may even be counterproductive in triggering negative reactions," the paper states. "Therefore, companies should consider how they can balance company needs and transparency to provide their 'workers' with a real sense of fairness and partnership."
The researchers advise platform companies to find ways for those using the platform "to participate democratically in decision algorithms and policy."
They also stress the need for human support at times when drivers need answers, pointing to Uber drivers in New York who said they switched to the competing Juno service because Uber's automated support emails contributed to feelings of dehumanization.
"Algorithmic management platform owners cannot expect both to be 'partners' with their workers and to keep their algorithms completely opaque," the paper says. ®