Tech giant Uber鈥檚 experiments with autonomous cars have been making a stir. It has been field testing a technology no one is certain will work. In the process, it has expressed disdain for the capabilities of human drivers. Who can blame it? Crashes kill about . Some 90 per cent are . That鈥檚 a fact, and I write this as an Uber driver.
But when the company, best known for its app that , made its robot cars available to passengers in San Francisco (all with a human chaperone behind the wheel), Uber faced a backlash. It pulled them out because it didn鈥檛 have the permits required, but also after the experiment started to get scary. Several of the cars ran red lights. , blaming human error and again pointing to the potential of autonomous systems to improve safety.
That wasn鈥檛 the only issue, though. The cars also , which Uber reportedly put down to a software issue. Why should we believe the cars won鈥檛 make such mistakes, or even worse ones, when completely autonomous? I鈥檓 not the only one thinking this. US customer rights champion so the public can better judge how the technology is performing. And for autonomous cars.
Advertisement
I can鈥檛 help pondering the central issue of safety while behind the wheel of my Prius on the way to my next passenger. Would a self-driving car be able to deal with all this snow? It is piled alongside the road, and the ploughs haven鈥檛 touched many side streets.
The Prius, a 2013 compact model, does pretty well in snow, but I chalk this up to years of winter driving in the US Northwest. There are so many variables in inclement conditions. It鈥檚 not just about what other drivers will do. It鈥檚 about the unpredictability of the surface beneath, and how to steer into a slide 鈥� the kind of instinct that develops over years.
Sensors vs sentient
Matt Sweeney, head of , says that have sensors to cope with the poor conditions I鈥檓 talking about. Radar units can bounce signals off the surface of the road underneath or some way ahead. This gives the computer data on what鈥檚 there and what鈥檚 coming up, even on busy roads. There are also optical cameras, LiDAR and ultrasonic detectors.
Are state-of-the-art sensors enough to win consumer confidence? I decided to supplement Uber鈥檚 experiments with one of my own 鈥� a simple survey to find out what my passengers think about driverless Ubers. The result is a small, but indicative picture of how confident people are in the technology.
Out of the 50 fares surveyed, 37 said they wouldn鈥檛 feel safe in a driverless Uber. Of course their responses could be skewed, given that they were talking to an Uber driver. At least in part, they could have said this to make me feel good. But they raised valid concerns. What if the cars get hacked? What if something really unexpected happens? How will a computer that has never dealt with it before adapt? Of the 37 doubters, 20 cited the San Francisco trial.
OK, so my survey was not scientific, but it tallies roughly with one by market research firm ReportLinker, which found that wouldn鈥檛 feel safe in a driverless car.
Apparently, those looking for a ride do feel safe enough to summon a human driver, because today鈥檚 human-driven Uber is valued at .
When it comes to driverless taxis and safety, Uber CEO Travis Kalanick and co have some convincing to do. The gig鈥檚 not quite over yet, fellow humans.
