Thanks to complaints from Uber drivers, who were beginning to suspect that the ridesharing company was charging customers more with "upfront pricing" but not paying drivers more in turn, on Friday Uber admitted in a Bloomberg report to using AI to find the upper limit of what people are willing to pay for a ride based on their route in 14 cities.
The revelation essentially confirms the suspicions of critics like the University of Washington's Ryan Calo and Alex Rosenblat of the Data & Society Research Institute. Together, they wrote a scathing paper that warned Uber may use vast amounts of customer data to act in a predatory manner—for example, price gouging based on your income level or other circumstances.
This isn't quite what Uber is doing with what it calls "route-based pricing," but the end result may be the same. Basically, Uber uses all the data it has on customer behaviour along particular routes in a given city to serve people different fares based on where they're going. To get an idea of what this means in the context of income inequality, here's an except from the Bloomberg report:
[Daniel Graf, Uber's head of product] said the company applies machine-learning techniques to estimate how much groups of customers are willing to shell out for a ride. Uber calculates riders' propensity for paying a higher price for a particular route at a certain time of day. For instance, someone traveling from a wealthy neighborhood to another tony spot might be asked to pay more than another person heading to a poorer part of town, even if demand, traffic and distance are the same.
What this means is that the wealthy may end up paying more for a ride on average if they're consistently travelling, say, from work in a business district to their swanky neighbourhood. On the flipside, fares for the poor may also be jammed up to the limit of what they're willing to pay for a ride if they consistently travel to and from a poor area. It's important to note that Uber isn't looking at individual customers' circumstances, the company claims, but group statistics.
"Uber has narrated a lot of its expansion with the idea that it's going to supplement public transit options," Rosenblat said in a phone call. "If the turnaround is basically that they can charge people more for living in an underserved area, then that contradicts Uber's earlier narrative."
Uber has forged partnerships to supplement public transit in several US cities, and in the small Canadian town of Innisfil, Ontario, the company is now providing the sole publicly-subsidized transit option. When transit is run by cities, however, fares are usually decided by balancing an affordable rate for riders with the cost of operating the service, not charging people as much as possible for a profit.
"What guarantee do users have that they won't be individually selected to receive a higher price?"
"We price routes differently based on our understanding of riders' choices so we can serve more people in more places at fares they can afford," an Uber spokesperson said in an emailed statement. "Riders will always know the cost of a trip before requesting a ride, and drivers will earn consistently for the work they perform with full transparency into what a rider pays and what Uber makes on every trip."
Given some of Uber's past transgressions, like the now-infamous "greyball tool" that used the company's data on riders to blacklist law enforcement, Rosenblat said that the risk is there for Uber to use the AI behind route-based pricing to set fares for individual riders, instead of merely everybody who happens to travel along a particular route.
"Uber certainly has a track record for using personally identifiable information to profile someone for their capacity as a user," Rosenblat said. "What guarantee do users have that they won't be individually selected to receive a higher price?"
Right now, that guarantee is just Uber's word.
Subscribe to Science Solved It , Motherboard's new show about the greatest mysteries that were solved by science.