Surveillance pricing exposed at Uber

Last week, The Capitol Forum published an excellent piece on how Uber infers how sensitive passengers are to price. Drawing on internal documents, it reports that Uber built a
machine-learning system that uses riders’ personal data to tailor discounts and funds those discounts by raising fares on certain routes determined to be price inelastic. The article is paywalled, but it’s quality journalism well worth supporting.
Think about a trip to the airport when you have a flight to catch, or a ride home late at night when public transport has stopped. How much more might you pay when you have few alternatives?
This kind of profiling has been around for years. As early as 2020, Ola Cabs’ privacy policy said it could use the make, model and version of a customer’s phone to infer their economic status and tailor offers and services accordingly.
Last month, Grab’s engineering team described an 'invention' proposing to use residential property values to group users by inferred affluence and gauge their price sensitivity. Grab says the solution has never been implemented or tested in a live market. (Doest thou protest too much?) The proposal itself deserves scrutiny. Why should the estimated value of homes in your area help determine which discounts you’re offered? Will people in poorer neighbourhoods be offered lower prices? Are platforms about to become the modern incarnation of Robin Hood? I doubt it.
Now think about the worker on the other side of the app. A passenger with few alternatives may pay more. A driver who has already spent an hour waiting unpaid may accept less. What happens when the platform setting the pay can predict how little a driver is likely to accept?
This is why WIE International has launched the Dynamic Pay Claim to protect all workers from such abuse.


Comments