Long-thought as something of science fiction, driverless cars are now legal on the roads of the UK. As the country establishes itself as a leader in driverless car technology, some questions still need answers – one of the most important being, who is responsible in a driverless car crash?
How will driverless cars change the UK’s roads?
Human error is blamed for 90% of all road traffic accidents, giving driverless cars (which use computers and sensors to drive, rather than human input) the potential to make driving much safer than it is manually. This is because these cars are able to predict their surroundings much more effectively than humans by using panoramic cameras, laser imaging, and radars (amongst other things) to get a 360° view of what’s going on around them. So confident is the government in these cars’ ability to predict the road that they’re lifting a ban on tailgating to allow driverless cars to drive closer together and prevent gridlock at peak times.
This ability to be hyper-aware of their surroundings means that driverless cars, according to Matthew Strebe, CEO of Connetic, will be near impossible to crash. “Two driverless cars will never hit each other, unless a human has caused an accident that they are both involved in,” he says, before continuing to say that “[a driverless car] will rarely if ever be found to be ‘at fault’ in an accident.”
But what happens when an accident is unavoidable – what does the car decide to do then?
Could there be a driverless car crash?
The number of road collisions will more than likely fall as driverless cars become more prevalent on our roads, though the issue of what a driverless car will do in a collision situation is still undecided.
Some collisions are unavoidable. A car could pull out in front of you, someone could cut across your lane or a car could take a red light into your path. In cases like this, when a collision is imminent and your only option is to hit another vehicle, humans make an emergency decision. The driverless car, however, will need to have something written into its programming telling it what to do in these situations – giving the car’s manufacturers and programmers an ethical conundrum. Which car do they make it hit?
A larger car is more likely to take the impact better than a small car, so it may seem favourable to program a driverless car to crash into the biggest car. However this would mean that larger cars would effectively be penalised for their size. Likewise smaller car owners could drive dangerously, knowing that they’re statistically less likely to be chosen. The same example can be applied across many aspects of driving – a cyclist with a helmet is more likely to survive than a cyclist without a helmet, however this could lead to fewer cyclists wearing helmets.
This selection bias is something that will have to be written into the programming of cars before they roll out the dealerships. Whether car companies opt for the least damage, or use a random generator to make collisions, is still to be decided. It’s just one of a number of ethical questions associated with taking the human element away from driving.
Who’s to blame if a crash does occur?
Whilst assigning blame in a road collision between two human drivers is already difficult enough, deciding who is at fault in a collision with a driverless car is much harder. If it was the fault of the driverless car, is it the car itself which is to blame, the passenger, the manufacturer of the car, the people who programmed the car to make that decision, the government or council for allowing the car onto the road, or another party completely?
The car itself is unlikely to be at fault in a crash, as it will be operating on rules it was given by the programmer. The programming of the car will need to be exhaustive, giving the car instructions on what to do in every possible situation – including what to do in a crash.
As the programming will need to be signed off by the company that makes the car, then responsibility could fall to the manufacturer.
The manufacturer will be, as it currently is, responsible for everything it produces. We’ve seen it happen time and again – a manufacturer releases a car, they find a fault with it and have to recall all affected cars. Should a manufacturer find something wrong with the programming of a driverless car, it will ultimately be the manufacturer who bears the burden – unless of course, it’s the sensor’s fault.
The sensor manufacturers
Major car manufacturers may rely on external manufacturers to provide integral parts like sensors, radars and cameras. If a bug in one of these components causes a malfunction, then responsibility would fall on the company which made the malfunctioning equipment.
The human driver
It’s highly unlikely that a driverless car will cause a crash as they are incredibly hi-tech pieces of equipment that are able to predict other road users much more accurately than humans. They will be able to see and understand when a vehicle is slowing, indicating, setting off or pulling out and adjust accordingly. In a collision between a human driver and a driverless car, it will be both statistically and factually more likely that the human driver is responsible for the collision.
The human driver being at fault is a sentiment echoed by charity Road Peace “whilst casualties and collisions are expected to decrease, as long as some motor vehicles are being manually driven, collisions will continue.”
What’s next for driverless cars?
For the time being, driverless cars will retain the ability to have human input. There is still a steering wheel and pedals, just like in a normal car, and the passenger can take control of the vehicle by moving the steering wheel or pressing a pedal. Therefore, if the human passenger feels that the car hasn’t recognised a hazard, they can manually take action.
Accidents are predicted to drop dramatically though, making the roads safer. It’s not just on the roads that driverless cars are excelling however – they’re also showing they can push a car’s limits on the track. Stanford University used an upgraded Audi TTS on the Thunderhill Raceway in California to set a time 0.4 seconds faster than the track’s CEO – proof, if needed, that a driverless car is able to interact with the road and itself better than its human counterparts.
It’s this ability to judge performance and timing better than humans which will help lead to much higher fuel efficiency, reduced journey times and safer roads. Driverless cars will be able to interact with each other coming up to junctions and traffic lights meaning that, rather than stopping, they will be able to adjust their speeds to fit through gaps. This is fine for cars, but what about cyclists and pedestrians?
Without a stop in traffic, it could become much harder for pedestrians to cross the road. The new relationship between pedestrian and car is something that is being tested in the UK trials – something Simon Williams, spokesperson for RAC, thinks is a step in the right direction. “The announcement that testing has started is an important first step along the road to us seeing driverless cars become common place in the UK. But there is a significant way to go and the prospect of driverless vehicles entering service before the end of this decade looks ambitious.”
Many of the issues surrounding driverless cars are still undecided. As with any technology, it will doubtless be an ever-changing and adapting process, similar to the aviation industry. And, just like the aviation industry, driverless cars will soon become as commonplace and accepted as autopilot on planes – just another part of an already well-automated method of travelling in the 21st century. Until then, a lot remains to be seen about this very new, futuristic technology.