What does the Uber crash mean for driverless cars?

Written by Goran Nastic

Uber crash may fast track the issue of fault and liability with driverless cars as an autonomous future looms.

Late on Sunday, a self-driving test car from Uber Technologies hit and killed a pedestrian in Tempe, Arizona, a state promoting itself as a testbed for such technology. It is the first such fatality involving an autonomous test vehicle and a pedestrian.

As expected, the crash has provoked a response from consumer-safety advocates and will likely lead to closer regulatory scrutiny. Jason Levine, executive director of the Center for Auto Safety, was quoted as saying in Bloomberg that the accident will set consumer confidence in the technology “back years if not decades”.

The sad reality is that the incident was inevitable, which the automotive industry feared, even as one of the key long-term goals of driverless cars is to dramatically cut all road incidents and accidents. It is estimated that over 3,300 people die every day on streets every worldwide as a result of human error. Incidentally, it is thought the first fatality involving a car was in 1896, when a prototype vehicle ran over Bridget Driscoll in London moving at 6km/h.

Uber CEO Dara Khosrowshahi has publically apologised in a tweet. Uber is spending hundreds of millions of dollars building its own self-driving cars. According to reports, loss-making company spent $1.17 billion on R&D in 2017, much of that on autonomous-driving technology. There are also reports that Uber may sell its technology to Toyota, a car maker it has worked with.

Toyota, Alphabet’s Waymo unit, GM, BMW and Ford are among those planning for an autonomous future. Most of these have a timeframe of early 2020s for such vehicles to be in commercial operation on the roads, if not perhaps at a large scale.

These are still early days, and computer algorithms face a steep learning curve, in some ways comparable to humans, to perfect how they respond to urban environments. There the much discussed scenario of what the algorithm does – or what it is programmed to do – in case an accident is unavoidable. Does it run over an elderly person, pregnant woman or someone else entirely?

For now, companies such as Uber will be held responsible when there are collisions, while the insurance industry is also grappling with car insurance itself that comes as complicated software systems (and new forms of hardware) make their way into vehicles. In the longer term, insurers and the automotive industry will have to decide on the complex minefield of fault and liability. It is thought this will involve the manufacturer, but also the software coders, IT managers, engineers and human ‘drivers’.