The application for a patent on "Safe routing for navigation systems," filed on June 26 and reported on by Roadshow, describes features to let ride-hailing app users prioritize safety along with things like distance, cost and trip time. The software could take several factors into consideration in calculating safety, including crime reports, the car's condition and how often the driver swerves.
One thing the technology wouldn't do is keep someone from getting picked up or dropped off in a given neighborhood. Uber spells that out in the application, possibly to head off charges of racial discrimination that have been raised against both the taxi and ride-hailing industries.
But it could use several possible sources to determine the relative "safety" of different areas that a route might go through. If a rider put a high priority on safety, the service might find a route that avoided those areas even though the trip would take longer.
Those sources could include crime statistics, news databases, academic databases of reports of violent conflicts in certain locations, and satellite imagery showing things like storms, fires and the effects of earthquakes. But the application also mentions some factors that can be a little more squishy, like event-listing websites -- "Particular types of events may be known or believed to be less safe" -- and social media, "which can be analyzed to identify (for example) sentiment about events, companies, and the like."
Uber drivers might feel the biggest impact from the system Uber wants to patent.
It would allow the app to determine the safety of a ride partly by factoring in telematics data from each vehicle and driver. That could include things like speed, amount of swerving and amount of braking for individual trips. This data could be used to identify dangerous parts of a road, since traffic and other factors can affect driving. But it could also be used to determine how safe it is to ride with a particular driver.
Riders can already slam their drivers by rating them on a scale of 1 to 5 after the ride. But that score reflects their experience as a whole, including whether the pickup was on time, what route the driver took and even whether the driver and customer got in a fight.
What Uber seems to have come up with is a system that could automatically log what looks like less-safe driving whether riders object or not. (If a driver has a history of conflict with riders, that could also be factored in.) Plugged into customer profiles that prioritize safety, that might affect whether a driver gets offered some rides.
Even a great driver might end up at a disadvantage if their car is older or has fewer safety features than others. The technology Uber proposes could use vehicle make, model, year and condition to gauge safety.
"Certain vehicles are safer than others because some vehicles are more reliable, have better performance, are more structurally secure, have advanced safety features, and/or are less likely targets of theft," according to the application.
Uber left its options open for how it might implement a safety preference. In addition to users specifically weighting their ride request toward safety, the company could build specific ride profiles with a preference for safety built in. Then riders could select a ride type, like "Family," that would automatically book a ride taking some of these factors into consideration.
— Stephen Lawson is a freelance writer based in San Francisco. Follow him on Twitter @sdlawsonmedia.