Self-driving cars developed by Google and others are likely to be technically superior before long, but their road to the mainstream will not be simple
It would be easy to believe that robot cars are powering inevitably onto our roads. Google’s self-driving vehicles have totted up more than 1.2 million miles – 90 years worth of driving experience – on California’s streets since 2009. Various countries and US states are bending over backwards to accommodate driverless car research. And the car industry isputting billions of pounds into research in an attempt to keep up with Silicon Valley’s deep-pocketed entrants.
This optimism is well grounded. Driverless cars promise enormous benefits: a revolution in productivity, a near-zero level of road fatalities, faster travel and a seismic improvement in energy efficiency.
And cars, when you step back as much as possible to think about them, are pretty terrible: they kill 1.2 million people a year; they are expensive – the biggest investment many people make after a house; and they are horribly inefficient, sitting idle 96pc of the time.
Driverless car disciples have a near-Utopian vision of personal transport. Instead of owning our cars, we’ll rent. A constantly-moving swarm of autonomous vehicles will roam the streets, picking up one passenger soon after dropping off the last and ending parking forever. We’ll spend the time that we now waste on driving doing something worthy.
A lot of this picture, if not all of it, is very possible. Driverless cars are inevitable – the amount of money being thrown at them, along with the potential benefits, add up to a powerful argument. The technology, though it hasn’t been tested in a lot of environments, has improved rapidly, and should be ready to graduate from tests in controlled locations and suburbia to more challenging urban scenarios soon.
It is very possible that before long, we will have a fully-capable autonomous vehicle. But for driverless cars to become a reality, that will only be the beginning. There are several tests that they will have to pass before we’re ready to get rid of our steering wheels.
Handling the first driverless death
The first will be how society reacts to the first major accident. Public opinion on driverless cars at the moment can be described as “healthily sceptical”.According to a recent study from the University of Michigan’s Transportation Research Institute, two thirds of Americans would be “very concerned” to “moderately concerned” about riding in a fully-autonomous vehicle, and 96pc would want a steering wheel and pedals as a back-up.
This would be likely to change as self-driving cars become more visible, but it shows a level of initial unease. Even if the mood begins to change, and autonomous vehicles start to appear on our roads, at some point, one is inevitably going to be involved in a fatality. It probably won’t be the driverless car’s fault – in the more than a dozen accidents involving Google’s cars, human drivers have been responsible for all of them – but the association may be hard to shake off, even if statistics show driverless cars are significantly safer than human-driven ones.
Speaking of accidents, the ethical decisions made by driverless cars in life-or-death scenarios are going to draw immense scrutiny. In the event of an unavoidable crash, software may have to choose between saving a passenger’s life and a pedestrian’s.
Whatever choice a car makes will be a matter of constant debate. Should car manufacturers be allowed to sell models that give higher priority to passengers – they would surely be more popular among buyers? If not, who should decide a driverless car’s ethical programming?
How secure is a driverless car?
The third major issue will be hacking. As basic levels of automation– auto-braking, self-parking technology and electronic immobilisers – come to today’s cars, vehicles are becoming bigger targets for attackers. Fully driverless cars would give successful hackers significantly more control, potentially allowing them to use the vehicle as a weapon of kidnapping or murder.
For years, car manufacturers put no piece of connected technology more complicated than a radio in their products. As they race to bring driverless abilities to the market, we have to hope that security is given its proper attention.
Regulating cars off the road
A fourth test, and one of the more unknown quantities, is regulation. Governments have been keen to welcome research into driverless cars, and the economic activity it brings, but actually allowing them on the road could be something else entirely.
Huge parts of driving law will have to be rewritten for a driverless future. Who will pay in the event of an accident? The driver can’t be responsible, given that they were not controlling the car. If you point to the owner, then they are likely to blame the manufacturer. In this case, how is insurance going to be worked out?
A lot of powerful entrenched industries are also going to be shaken up by self-driving vehicles. Look at how London’s taxi drivers have responded to the arrival of Uber and imagine what would happen if they were undercut many times more by taking away the cost of a driver. Add public transport unions, truck drivers and logistics networks and there will be a powerful opposition.
And finally, but not insignificant, are our habits. The mass car ownership model is more than a century old. Cars are a status symbol and something to be proud of for many people. And drivers like to be in control, even if it is unsafe and time-consuming.
Even if the technology behind driverless cars is practically perfect, its road to the mainstream is likely to be far from smooth.