Self-driving Uber car-pedestrian death reveals how we’re nowhere close to Level 5 tech

Progress for a world of self-driving vehicles may have taken a step backward. On Sunday, March 18, the first report of a pedestrian fatality by a self-driving car occurred in Tempe, Ariz., after an autonomous Uber SUV struck a homeless woman.

At approximately 10 p.m. local time in Tempe, 49-year-old Elaine Herzberg was walking her bicycle across a street. At the same time, an autonomous 2017 Volvo XC90 owned by Uber was driving down the street at 40 mph.

The Uber driver did not see the woman and struck her. She was sent to a hospital, where she was pronounced dead.

The Uber driver did not appear to be “impaired.” He likely was not paying attention. Why?

The Uber Volvo XC90 was in autonomous mode during the incident.

Does the first pedestrian fatality of a self-driving vehicle mean anything or are some people extrapolating too much out of a single incident?

To start, autonomous vehicle manufacturers have a lot of data to hold onto when defending the safety of their technology. Millions of miles are on record during testing with very few incidents.

Data will also make the death of a pedestrian less surprising. From 2007 to 2016, pedestrian deaths increased by 27 percent while other categories of traffic deaths decreased by 14 percent.

According to a Governors Highway Safety Association report, five states each reported more than 100 pedestrian deaths: California, Florida, Texas, New York and – wait for it – Arizona. The first four states are the fourth largest states in terms of population. Arizona ranks 15th. In fact, Arizona had the highest rate of pedestrian deaths per resident population.

I’m not sure what is going on in Arizona, but if had to place a bet where the first pedestrian death by self-driving car would take place, Arizona would be easy money.

Furthermore, in this case, a human driver was behind the wheel. This helps the manufacturer because it implies this vehicle was never ready to be fully self-driving in the first place. The biggest argument for self-driving vehicles is the fact that more than 90 percent of traffic fatalities are the result of human error. Lawmakers and stakeholders cite that site ad nauseum in every single congressional hearing about autonomous technology.

In fact, the man who was killed while driving his autonomous Tesla was found not to be attentive while the car was in autopilot. All legal autonomous vehicles require a driver behind the wheel to take over. This leads us to the argument against autonomous technology…

Humans are not ready for autonomous vehicles.

Despite human error causing 93 percent of traffic fatalities, we will need to perfect Level 2 and 3 autonomous technology (requires driver attention) before even thinking about Level 4 or 5 technology which allows us to literally fall asleep at the wheel.

Both Volvo’s XC90 and Tesla’s Model S uses Level 2 autonomous technology. Both cars involve a fatal crash where the driver overestimated the car’s ability to drive itself. Despite these incidents, Volvo has announced plans to roll out a Level 4 vehicle by 2021. With Elon Musk’s super-lofty goals, Tesla is hoping to build a Level 5 vehicle in the same time frame. However, neither can manufacturer a completely safe Level 2 vehicle.

Even though the safety record of autonomous cars may be better on paper and in real life, the general public expects a certain level of perfection with fully self-driving vehicles. A recent Pew Research study reveals that more than half adults will not ride in a driverless vehicle. Nearly three-quarters worry about robots and computers performing many human jobs.

Due to the lack of efficiency with current Level 2 vehicles and the hesitance of people accepting self-driving cars, the future of sleeping at the wheel on your way home is probably much further away than lawmakers and industry stakeholders have been claiming.