In 2004 the US defence research organisation Darpa set cars the challenge of navigating a 150-mile desert course without drivers none of the cars finished and the two leaders only managed seven miles.
Ten years on and Google’s tricked out Priuses have travelled over 700,000 miles around California without a single accident.
And recently Google unveiled its very first custom-designed autonomous car which breaks the mould in several important ways. It only has a start button, for instance, not all the controls we normally associate with driving. And it looks less like a car and more like the sort of transport you would take in Disneyland.
It’s quite clear from what Google has revealed about how the sensors and algorithms work that enormous progress has been made, and despite the challenges still ahead it is now possible for forecast with confidence a time when autonomous cars could be ubiquitous.
And this could be truly revolutionary. Google’s design choice is I think intended to point out how different a self-driving car can be from what we think of as cars. The self-driving car could usher in a world where owning a car becomes completely unnecessary – we could simply summon a car from our smart phones and rent it for only the specific journey we need to make. Parking would be a thing of the past – the cars would simply drop you off and either go to their next job, or back to the charging station to recharge themselves.
Given that 96% of the time our cars are parked and doing nothing, a complete move to self-driving rented journeys would vastly reduce the number of cars required, saving resources and parking space at home, work and on city streets. There could be massive benefits to our urban environments with concrete making way to green space again.
This is to say nothing of the tremendous life-saving potential. The World Health Organisation estimates there were 1.24 million deaths caused by cars worldwide in 2010. The promise of autonomous cars, always awake at the wheel, never getting angry and frustrated, never getting drunk or texting at the wheel is of vast reductions in these numbers and all the suffering and economic loss which comes with them.
Given the enormous benefits in economic, resource and health terms surely we will see a world which quickly adopts this marvellous technology?
Sadly, I doubt it.
Although the technological hurdles are clearly eminently solvable and in a relatively short time frame, other difficulties remain. Firstly, the complex regulatory and legal issues which will need to be addressed are not trivial. How will be liability in an accident be handled, for instance.
The most tricky issue, though, is likely to be the libertarian arguments about the right to drive. According to the FBI in 2012 there were 8,855 total firearm-related homicides in the US. And yet the gun lobby continues to insist on the right to carry guns despite the carnage. Expect similar freedom-related arguments about the right to drive.
And the transition from driven to self-driving cars could have unexpected consequences. Self-driving cars are very good at avoiding hitting things, so good in fact that the behaviour of other road users may become more and more reckless and traffic speeds as a consequence may drop considerably as self-driving cars become more and more cautious. This may be true of pedestrians and cyclists too – again leading to slower and slower speeds.
As so often, the technology will likely be much more advanced that the soft issues around it.