CARnage?

Autonomous vehicles are coming to a road near you, and soon. The race to dispense with drivers is well and truly on. It will inevitably cost jobs, but will inevitably save lives, too.

It is reported that 94% of road accidents are caused by human error, and that these accidents kill 1.3 million people every year, leaving another 20-50 million injured and/or disabled. These are shocking statistics, and a terrible price to pay for the convenience of travel.

We spend a lot of our lives driving, too. The average American spends just over one hundred minutes driving, every day. Every day. This equates to over four years of the average American life. Four years, spent doing nothing but concentrating on not becoming another casualty of our global obsession with automobiles. Four years when we could be doing something else.

And then there is the cost to business and commerce. Many businesses employ a range of driving personnel – taxi firms, hauliers, couriers…the list goes on – and they account for 3% of the workforce.

The case for developing self-driving vehicles is compelling. And it seems that we are on an inexorable path. Perhaps that should be road. Billions of dollars and Euros are being poured into the pursuit of something that even twenty years ago was nothing more than a science-fiction pipe dream. Technology never stands still though, and the rise of AI, coupled with the era of big (soon to be huge) data means that driverless cars will be on a road near you in the next five years or so.

General Motors might be leading the way, but perhaps they are merely winning the PR battle. Waymo – with strong links to Google – has been developing the tech longer than anyone, and they are close to launching an autonomous ride-hailing vehicle. Daimler-Bosch, Ford, Volkswagen, BMW, an alliance of Renault, Nissan and Mitsubishi, as well as one between BMW, Intel and Mobileye, and another of Volvo, Ericsson, Autoliv and Zenuity are all investing heavily in their self-driving projects. Aptiv and PSA are two new players in the market. And of course, Uber and Tesla are making huge investments, too. When you have so many iconic, globally recognised organisations developing the technology, it seems inevitable that they will succeed. Some will be more successful than others, though, and it will be fascinating watching the battle for a market share. There may be some big name casualties who turn up too late to this party, too.

A problem with any new technology is that we often play catch-up with ethical and regulatory issues, but at least we have the advantage of seeing these developments coming. The Society of Automotive Engineers (SAE) devised a system of five levels of automation (six, if you include zero automation) that has been widely adopted as an industry standard.

Level one is already commonplace. The driver still drives the car almost entirely (“hands on”), but software that helps us park or controls speed under certain conditions can be found in many modern cars. Level two (“hands off”, although, somewhat confusingly often not literally) means that the car is self-driving, but requires constant human monitoring and intervention if necessary. There is no real suggestion that this is a workable option in reality; it is a stepping stone towards the completely autonomous goal.

Level three (“eyes off”) means that the driver does not have to monitor what the car is doing, but may be required to intervene within a specified time period. It’s nearly there! Level four (“mind off”) is all but there. The driver (okay, passenger, by now!) can have a nap if he/she wants to. A large whisky or two is he/she is so inclined. The vehicle is entirely autonomous with one caveat – it is geofenced, which means it can only operate within a certain geographic area. It can sometimes operate outside of this area in a traffic jam, but the vehicle is required to understand when it has reached the limit of its capabilities within these caveats and must abort the journey and be able to park safely and disengage unless taken over by a human driver.

And finally, we have the holy grail of level five (“steering wheel optional”). The vehicle is entirely autonomous under any road conditions, and requires no input at all from a human.

It is hard to know exactly when the first production line, fully autonomous vehicles will be available to the general public, but the more optimistic predictions suggest it could be as little as just two to three years away.

The legalities are going to be very complicated. Lawmakers are already trying to anticipate the onset of self-driving cars, but it seems an impossible task to legislate for every potential eventuality. What will happen when such a vehicle malfunctions, and someone dies? Where will the responsibility lie? Who will be prosecuted and what charges will they face?

Car insurance is going to get very complicated, that’s a certainty.

From a statistical perspective, the more robotically operated cars on the roads, the safer they should be, and less people will die in accidents. There are two problems here though. Firstly, deaths will probably go down, but it is impossible to know who has been saved a premature death. We don’t have two different realities to compare, so the lives that are saved will never know that they have been saved. The technology will not get the credit it deserves, but you can guarantee it will be quickly condemned for the people that still die.

Secondly, 90% of road deaths occur in low and middle-incomed countries, even though they have less than 50% of the world’s vehicles. These countries will inevitably be the slowest to feel the impact of driverless cars, meaning that global road deaths will remain north of one million for decades to come.

And what of hackers? Self-driving vehicles are likely to be a target for those that take pleasure from interfering in technology and the lives of others. There have already been plenty of tabloid future-shock stories, which by their very nature look to exaggerate and sensationalise, but the potential damage that could be caused by the e-hijacking of an articulated lorry (just to make it as dramatic as possible) could be truly appalling.

Of course, the coders (and the power that these people are beginning to wield is a blog within itself) will do their very best to stop this happening…but this will be a battle that will crank up as the proportion of autonomous vehicles on our roads increases.

It’s all fascinating. And more than a little bit scary. We appear to be living in interesting times.

I think they’re about to get very interesting.