I regularly find myself justifying autonomous cars in idle conversation. Often, my defence of the robots leads to an accusation that I don’t love driving, or that I’m trying to invent Skynet. Nothing could be further from the truth. I really love driving*.
One of the most popular watercooler discussions of the moment is the ‘moral’ choice that a self-driving car’s software will be forced to make when confronted with an impossible decision; mount the kerb and hit two children? Continue ahead and collide with an SUV carrying seven nuns? Who is responsible for the decision? What heartless developer will write the algorithm that decides which family to bereave?
But this is sensationalism and grossly misrepresents both the probability of this situation arising and the enormous social good that self-driving cars will bring globally. I’d like, in this post, to consider some of the facts about autonomous cars and then stretch those facts into some fantastic fortune telling.
THE REAL COST OF CAR OWNERSHIP
In the UK, in the year ending March 2015, 1,740 people were killed in road traffic accidents. In the same period 23,570 people were killed or seriously injured, amongst 186,060 reported road casualties of all severities.
If those figures are grotesque, consider that the UK has some of the safest roads in the world. Worldwide, the WHO claim that about 1.25 million people die each year as a result of road traffic accidents, which are also the leading cause of death among young people (aged 15–29 years). The WHO also report that half of those dying on the world’s roads are “vulnerable road users” — pedestrians, cyclists and motorcyclists.
It’s not all quite such grim reading; there is a global decline in fatal accidents despite growing car ownership, with more countries reporting decreasing numbers of road deaths than those reporting increases. The WHO credits this decline to “improving legislation, enforcement, and making roads and vehicles safer.”
And this will be the very crux of my argument: in the future autonomous cars will not only be safer than human drivers, but they will be so much safer that it will be considered as morally reprehensible to drive yourself in a city as it is to smoke in a restaurant.
SAFETY AND PERCEPTION
You and I, dear reader, are special. Let us accept that we are above average drivers — more alert, more aware, more experienced. We’re above the centre of the bell curve for safe and responsible driving. But of course, half of the driving population is below average, which means that we have to also accept that there are a large number of people who are more likely than you and I to cause an accident.
Let’s think about these other drivers for a moment. How many times have you witnessed cretinous disrespect for other road users? Not a journey passes where I don’t observe random lane changes, a mad dash across the last yards of a motorway slip road or a driver fiddling with something more interesting than the road ahead: Facebook, lipstick, in one case an actual plate of food, complete with knife and fork.
Sadly this is not just the case with casual drivers. In the vitriol laden air of my car, the words ‘professional driver’ are now an epithet, barbed and laced with venom. From my experience driving in and around London, you can very nearly guarantee that a Toyota Prius, wearing a private hire badge is going to display some potentially murderous behaviour. So, if even the professionals are lethal, what hope is there to keep lowering those accident rates?
THE EVIDENCE FOR ROBOT CARS
Google’s autonomous cars have driven nearly 2 million miles and so far been involved in twelve accidents. Google has accepted partial responsibility for one accident, the bus collision in Mountain View which recently made headlines. The other eleven accidents were caused by other vehicles; most commonly with the Google car being rear-ended by other drivers. This puts Google at an accident rate of roughly 0.64 accidents per 100,000 miles driven, where the US average is reported as 0.38 per 100,000 miles.
Whilst these figures initially look stacked against Google’s experiment, it’s worth considering that the US National Highway Traffic Safety Administration (NHTSA) who provide these figures also claim that 54 percent of property damage accidents aren’t ever reported. This starts to make a stronger case for those Google numbers, which are still a long way from perfect. My number, as it happens, is much higher than Google, at 1.0 (2 accidents over about 200,000 miles in 23 years).
Let’s not forget that just like teenage drivers, today’s autonomous cars are inexperienced and will improve faster than we expect. The car industry (well, Volvo mainly) has an excellent track record of introducing innovative safety technology over the years — seatbelts, ABS, airbags and crumple zones through to today’s collision avoidance systems — and there is no reason to expect that they won’t do a similarly good job with full automation.
Google’s teenage autobots are learning not just from their highway driving, but also from their experiences in the lab, running over 3 million miles of simulated driving every day, which feeds the machine learning algorithms responsible for training the cars.
Tesla takes this learning a step further, using the worldwide deployed fleet of customer vehicles as a mobile lab, feeding back information on how their cars are driving and being driven. This is a company, let’s not forget, who launched their self-driving car software, Autopilot, as an over-the-air update. Tesla’s fleet, as Elon Musk himself explains, “operates as a network. When one car learns something, they all learn it”.
Outside of these digital upstarts, consider that VW, prior to their ‘creative engineering’ scandal, held the largest R&D budget of any company in the world from 2012–2014(significantly edging out powerhouses like Intel, Amazon, Google and Roche). Toyota, GM and Honda also make the top 20 R&D spenders regularly, and you have to expect that in the face of Google and Tesla, significant dollars are going into the exploration of safety through autonomy.
A WORLD WITH NO DRIVERS
While passenger car experiments are more immediately apparent, Mercedes also hit the news in the UK recently as the catchily named “Mercedes-Benz Future Truck 2025” is planned to start trials on the M6 motorway in Cumbria. A convoy of robo-trucks, ten astern, and controlled by a driver in the lead truck, is set to proceed down the motorway with vanishingly small gaps between the rigs. Unsurprisingly, the benefits are being touted as safety, economy and reduced congestion.
But we have to be honest and consider the direction that the needle is pointing. The convoy of Mercedes trucks is still piloted by a human. If one car is safer being driven by software, it stands to reason that ten trucks are, well, safer being driven by software. The driver in this convoy is present to make the public comfortable and appease legislators rather than to be a long term fixture.
So, here’s the startling reality that autonomous vehicles will usher in — that wherever possible, the moral imperative dictates that fallible humans should be replaced by autonomous vehicles. Not just in our own cars, but also in the fleets of trucks, taxis, ambulances, fire appliances, farm machinery and delivery vehicles on our roads. Wherever the job can be more safely and cheaply completed by software, pure Kantian reason dictates that it must be, for the good of humanity itself.
This will lead to a sudden trend which may seem distant now but that will unfurl in the developed world startlingly quickly. As the technology behind automating vehicles becomes commonplace and the price drops (as in the case of the $35,000 Tesla Model 3), affordable robot cars will become the norm. Not only has Elon Musk argued that laws may eventually prohibit human drivers, but Uber’s rambunctious CEO, Travis Kalanick, has forecast a time that it’s cheaper to use an UberX than to own a car.
Kalanick (a man so competitive that he’s quoted as saying “You can’t win 51 to 49. You have to win 98 to 2”) has to be looking past the expensive and fallible humans he has driving his logo around. Even before recent rumours emerged that Uber had made a deal to buy 100,000 autonomous limos from Mercedes, the obvious choice for a company as brand obsessed as Uber is to remove the one thing that can really vary from ride to ride — the driver.
An Uber with no driver may well be a better Uber — I was once asked by the (professional) driver of an Uber taking me to Heathrow airport what the national speed limit was — but the robo-Uber will also work without breaks, be polite to passengers, drive safely and most importantly do exactly what Uber requires of it. It’s hard to believe, when applying cold reason to the problem, that Uber won’t move to a driverless future. The only question is one of “how fast”, not “if”.
TRANSPORT AS A SERVICE
Stretching the facts to breaking point, our society is fast becoming one of consumption rather than ownership. The tropes of Blockbuster giving way to Netflix and wallets full of painstakingly procured CDs ceding to Spotify are obvious. Even in business Amazon and Microsoft’s cloud computing divisions, where physical servers give way to compute power available by the minute, threaten the core business models of Dell and HP.
No longer do people (or businesses) want the capital cost of ownership, but prefer the instant and omnipresent ability to consume on demand. I frequently use Uber as an example when talking through new product ideas. Distilled down, Uber is a ‘rescue me’ button, not a taxi service. It provides a safety blanket that allows you to go anywhere you like and still get safely home. Uber then has superseded the simple concept of ‘transport’ and already become TaaS.
When Uber forges ahead then, it won’t be long before others follow. Different business models, probably from the manufacturers themselves, will start to sprout. Imagine if instead of choosing an Uber you choose a lifelong service of cars on demand from VW, Toyota or Tesla. Instead of the massive capital outlay of buying a car, you pay as you go, and the car drives itself to your door when you need it — an SUV when you’re going on a trip with the family, or a limo when you’re impressing friends. Choosing how you arrive at your destination will be an app away.
This potentially has massive ramifications, and the 300,000 licenced private hire drivers in the UK who may soon be facing unemployment are just the very start. If cars become a service, the potential impact on the wider car industry as a whole will be enormous.
As more people become comfortable with cars that magically arrive to chauffeur them on demand, there will be a commensurate drop in the number of cars on the roads — or more specifically, parked in driveways. Congestion may not decrease, but the number of wastefully unused vehicles will. Driveways will be reclaimed, and equally as suddenly, the forecourts of second hand car dealerships across the country will start seeing their stock — and their customer numbers — plummet.
Car manufacturers themselves will be be forced to adjust from a model of franchised dealer networks and profitable leasing schemes to a pay-as-you-go model. Success for mid-21st century car manufacturers will be as reliant on the digital experience provided by smartphone apps as on manufacturing integrity.
Meanwhile, the century old second-hand car sales and service industry is going to be hit hard. Niche and specialist retailers, selling quaintly old fashioned antiques which still consume fossil fuels and use manual controls, will remain. School run people carriers and train station dump cars on the other hand will go the way of the dodo, fundamentally curtailing the customer base of large car supermarkets and small, family car businesses.
CATERHAMS, TESLAS AND TOYOTAS
In this fantastic and simplified model of the future, I don’t suggest that everyone will suddenly subscribe to the new world order. The sound of a naturally aspirated, petrol snorting GM V8 will still excite and motivate diminishingly small corners of the internet. Equally, there will still be those who are motivated in their choice of vehicle by demonstrations of wealth or status, or just a demand for luxury.
I give you then the simplified three-archetype model of the new car world. For the majority of people the ease of simply having a clean, reliable and functional vehicle at their disposal will be paramount. These people are pragmatists, who see cars as transport and probably today drive a Ford Galaxy, Nissan Micra or a Ford Fiesta (currently the UK’s best selling car). These are the prime candidates for Transport as a Service, and will make up the majority of the population.
A smaller group will still consider a vehicle as a sign of status and ownership a rite of passage. For these people, who may currently drive BMWs, Range Rovers, Aston Martins or Escalades, the car is not a pragmatic purchase. The car is a goal and an achievement, outwardly demonstrating their position and their personality. This group will still be tempted by ownership, but a large number will prefer to use these types of vehicles on demand. With a relationship not dissimilar to today’s ‘supercar clubs’, they will even be able to choose more luxurious and prestige brands because the diminished ownership cost will make them more achievable for special occasions.
The last group will be the most disrupted and emotional about this emerging economy. Predominantly male, likely of a certain age and income (let’s face it, me), they’ll believe in the romance of flat plane crank V8s and steering feel. They’ll be the people who currently obsess about handling, about ‘evoness’ and ‘feel’. Their world — and mine — will be a sadder, less colourful place without the cars that pack our imagination.
The salvation for this group will be in buying cars which light up more brightly — a world of cars built for the track or country roads. These are the Caterham drivers, the MX-5, Lotus Elise and Porsche GT3 owners, for whom the tactile joy of driving eclipses the necessity of A to B. But, this group will be the smallest, the smokers standing outside the pub. The driving joy of this group will require a new industry of non-autonomous roads and custom made tracks to exploit the cars we love in the way that they were meant to be used. The future isn’t particularly bright for us — but neither, in an era of speed enforcement cameras and increasingly capable vehicles — is the present.
Autonomous cars won’t forget the national speed limit. They won’t doze off. They won’t be distracted by family rows. They won’t miss the spectacle of wonderful scenery. They won’t get angry, drunk or morally indignant. Their male ego won’t be called into question by other robot cars.
They will always check that they are in good health before setting off. They will be able to see in the dark, and the rain, and the snow. They won’t take unnecessary risks to be on time for a meeting. They won’t have bad days — or if they do, they’ll just pull over and refuse to drive.
IN AN INSANE WORLD, IT WAS THE SANEST CHOICE.
The future of autonomous cars is one where travel will be an opportunity to catch up on work, talk to your family, ignore them with sound-cancelling headphones or check what restaurant TripAdvisor recommends in the next town. Autonomous cars will wait at the pub to drive you home and still get up early the next morning to take your kids to swimming lessons.
Crucially autonomous cars will do a better job at driving safely and economically than a human. By communicating with other cars around them, traffic will flow more smoothly, decreasing journey times and lowering the stress of travel. Fewer people — a lot fewer — will die, or be hurt in car accidents. For the largest majority of the population, the best solution is to put our trust in autonomous, on-demand cars and to enjoy how easy are lives have become without the stress of actual driving.
The challenges to today’s automotive industry and transport infrastructure will be massive, with professional drivers in taxis, coaches and trucks being suddenly surplus to requirements and forecourts of car lots across the country standing empty. Even plane and train travel will feel the impact, where transport as a service makes the private car the most convenient public transport system.
Perhaps, though, we’ll see the rebirth of the continental roadtrip, where the joy of the journey is greater than the thrill of the destination. Maybe a new economy of driving for pleasure will explode, with manufacturers and ‘experience’ companies providing for the last generation of petrolheads.
Whatever happens, make no mistake that the robots are coming, and they’re better than us.
* I love cars. I grew up in the car industry. My grandfather’s business, originally repairing lawn mowers, became a car dealership. My dad, a mechanic who became a reluctant car dealership owner, was happiest when he was recovering broken down cars. My brother sells the remarkable Caterham Seven. I’ve driven on many of the best alpine passes in Europe, the fabled Nurburgring and several UK race tracks. But this is a post about technology and perspective, not my love of super unleaded.