Will we ever have driverless cars? In my #AskAFuturist thread on Twitter, Ian (@ianeditz) asked a rather more negative version of this question: “Why will we never have driverless cars?” Ian is a sceptic when it comes to the idea of autonomous vehicles. In this post I will explain why he is right to be sceptical, but ultimately why cars will indeed drive themselves.
The case for scepticism
A truly driverless car, or autonomous vehicle, must be able to pilot itself in all conditions, with or without passengers. This means every street scenario, from a four-lane highway to the narrowest and busiest city street. And it means every weather situation, from three feet of snow to the glaring sunlight on a mid-summer’s morning. If you think this is simple, then you underestimate the incredible capabilities of the human brain, body and senses. We are unbelievably good information processors. We filter huge numbers of signals from the noise around us and react constantly with small adjustments to our speed, road position and more. Replicating and even beating human capability in controlled road conditions is within the bounds of current technological capability. But doing so in a range of conditions to which humans so readily adapt, is much, much harder.
So, before we can have truly driverless cars there is first of all a technological challenge to overcome. Can we equip the vehicles with an array of sensors, and a digital brain to process their output, that is better than humans?
Better than human
The key word in that last sentence is ‘better’. Autonomous vehicles don’t just have to be safer than humans, they have to be much safer. Perhaps an order of magnitude safer or even more. Because from their inception they are fighting a very natural fear in us: the fear of giving up control. It is scary to consider being in a vehicle that can move at life-threatening speeds and that is constantly making decisions about your safety and other people’s around you. Every time an autonomous vehicle fails and someone is injured or killed, the date by which most people will accept riding in an AV is pushed back. And sadly, this will happen many more times.
This is before we even get into the issues of security. What’s scarier than being in a rogue driverless car? Being in one that someone has deliberately hacked to cause harm. Right now car manufacturers are some way behind the curve on internet security. Lots of vehicles have shown themselves to be highly hackable. In fact the very nature of current vehicle electronics, using lots of components, both critical (e.g. the throttle) and non-critical (the stereo) interconnected over a relatively simple networking system, opens itself up to hacking.
Because of the incredible risk to life that autonomous vehicles represent, they will naturally be covered by extensive legislation. This will take a lot of time, as legislation does, however much ministers might like to announce programmes of support for the technology and the companies building the cars.
The legislation will undoubtedly require insurance cover that is somewhat different to what is required today. That too will take time for the insurance industry to organise. But who buys that insurance? Is it the owner of the vehicle? If so, what grounds do they have to be confident in the algorithms doing the piloting? So, is it the manufacturer of the car? Chances are that they may have bought in all or part of the software running the system, as well as most of the components that make up the sensors and systems on which it runs. So we have a complex chain of liabilities. This is true today, it just gets more complex when a human isn’t in charge of the vehicle.
It gets even more complex when you look at the trends in car ownership that are likely to be accelerated by autonomous vehicles. In short, most of us are likely to slowly move to greater reliance on fleet services like Uber rather than car ownership, especially in large cities. Car ownership seems to be losing prestige amongst young people who are learning to drive later and later. They are choosing to spend their money more on things to do than things to own (the subject of a future #AskAFuturist post). If a car can be at your door in minutes whenever you need it, and you don’t get to drive it anyway, why own one?
The case for the defence
Given all of these good reasons to be sceptical, you might rightly ask why I am so confident that we will – eventually – have self-driving cars. This comes down to four things: culture, cash, safety and good old-fashioned human laziness.
Culture: do we even want cars anymore?
What I mean by culture is that we care less about cars now than we did a generation ago. They aren’t the status symbol they once were. And in the context of a changing climate, owning a car – particularly a fast or extravagant one – is looking more and more like an unnecessary luxury, even an insult to your neighbours. I’m still a bit of a car nut but I recognise that fewer and fewer people share my passion. Self-driving cars allow us to get away from car ownership without requiring the large scale investment in public transport, cycling infrastructure, and city redesign that I would love to see, but do not see coming any time soon.
Cash: human drivers are expensive
The second point is more about the wider economy than cash in your pocket, though a subscription to a fleet service could be much cheaper than car ownership. Especially with economies of scale if most people go down that route. Human drivers cause accidents and traffic jams, which cost the economy and the tax payer an awful lot of money. I believe it is inevitable that self-driving cars will eventually be a lot safer than human drivers, saving us all a lot of money – and time.
Safety: computers don’t get distracted
That leads neatly to the third point: safety. Of course the greatest cost of failures by human drivers is not financial but the cost of lives, blighted and ended. Self driving cars will be, statistically, a minimum of ten times safer than humans. Probably at least 100 times safer. The lobbying power that will be brought to bear by road safety campaigners once a direct comparison is possible will be hard to resist.
Laziness: we like low-friction lives
Finally, there is good old fashioned human laziness. We have strived for a few million years to apply technology to take friction out of our lives. And what could be more appealing than a service that whisks you from one destination to another with barely any interaction required? Ultimately I think this leads us to overcome our fear.
So, will we ever have driverless cars?
Yes. But, it’s going to take a long time for all the reasons that Ian is right to be sceptical. Longer than people think. I don’t think the technology will truly be there for a ‘Class 5’ self-driving car that can operate in all environments and conditions until the end of this decade, and I think it will take a few years after that for all the legal and legislative wrinkles to be ironed out. Along the way, sadly, I’m sure more people will be killed by autonomous vehicles that aren’t quite there yet. There will be a public backlash against them. But ultimately, we will accept driverless cars because they make us richer and safer, and allow us to be lazier.
(A reminder at this point is worthwhile, that this is not what I want to be true but rather what I see happening. I can think of better alternatives but that’s not the question I’m trying to answer).