The whole subject of vehicle autonomy fascinates me.
I see lots of debate about when the technology will be “good enough” to become mainstream. We hear statistics about how many millions of test miles have been travelled without incident in realistic environments. Then, we also hear about the tragic incidents where people lose their lives because of autonomous vehicles not getting it right.
When David Clark revealed the findings of an impromptu survey he ran on LinkedIn recently, nobody thought true, full, driverless automation was going to be achieved in the next 5 years, 60% thought it would happen in 6 – 10 years, and everyone else thought it would take longer.
So, what does “good enough” mean? Is it enough that the occurrence of incidents is already claimed to be statistically lower than incidents caused by human drivers?
Because if that is true, and they are already “better” drivers than humans, why isn’t autonomy already more widespread?
There are other things at play, clearly. There’s probably a psychological hurdle for us to overcome around fault and blame. When accidents happen in cars driven by humans, we can, at least, process the harsh truth that humans are flawed, make mistakes, get distracted, make bad decisions, and so on. But for many people, to accept that a machine is allowed to be flawed screams “it’s not ready!”.
Assuming we can move past that (because, arguably, the technology will never be perfect), we start to think about other implications of autonomy. The areas of our lives and businesses that could be touched by it are far reaching.
Let’s start with insurance, ownership and liability. It’s seems obvious, that if Level 5 autonomy becomes a mainstream reality, then by definition the manufacturers of the vehicle(s) would have to assume liability for any accidents. Which vehicle, though, they’re both almost infallible, right? Or at least, they’ve both been approved in some way for fitness-for-purpose, and probably utilising a lot of the same technology and data.
Couple that with what it could do to the ownership model. If a vehicle is entirely autonomous, and we no longer get the experience of driving, will we be as bothered about owning a car? If it can drive itself, maybe I won’t bother owning one, and instead will just summon one that suits my needs on any given day. A fairly utilitarian box for my daily commute, and a more luxurious model for a longer journey at the weekend, up to my ears with entertainment systems and comfortable seating, perhaps? Apparently, owned cars are already only utilised for 5% of the time. The rest of the time they’re parked.
So, if I don’t own the car, and I’m not liable for any accidents, I don’t need insurance, surely. Does the fleet owner need insurance, because they can’t influence the driving success of the car either? Or is it just the manufacturer that needs to make provision for paying out for liability claims, which will happen infrequently?
If that’s the case, do they need to give any premium to an insurance company at all? Insurance is based on an assessment of risk, where risk variability typically relates to the human driver – autonomy levels the playing field. Presumably premiums would be high to insure a global fleet of millions of cars, and the money is gone whether they have to settle a claim or not. Perhaps instead the manufacturer will just budget and make provision for claims based on the expected error rates in the autonomous capability?
Of course, to get to true autonomy requiring no human intervention at any time, we have to pass through the tricky hybrid state of increasing automation and reducing intervention, but at what stage does liability stop sitting with the human that could intervene? Might that be the thing that slows down the journey towards Level 5?
So, maybe motor insurance as we know it becomes a thing of the past, and companies either keep reserves themselves or they buy liability insurance instead?
Then, we go further… I learned to drive in a car that had winders to put windows up and down, mechanical handbrakes, no ABS, no airbags, no power steering, no sat nav, no lane assistance, no intelligent braking, no cruise, no limiter, no… (the list is almost endless).
I loved driving those cars. Even modern, non-autonomous cars are fun to drive. Full autonomy makes driving a commodity experience. But in a world, say, 20 years hence, when autonomous cars are in the vast majority, humans driving their own cars suddenly become the weak link in the safety / liability equation. What will it cost me to insure myself to drive a non-autonomous car (assuming it is even legal) in an environment almost exclusively populated by robots? If there’s an incident, won’t it almost certainly be the my fault?
Perhaps I’ll only be allowed to drive a non-autonomous car at a track day, and I’ll have to get my car there in an autonomous carrier!
What about the manufacturer’s own business? If private individuals buy fewer cars, and instead consume on-demand services, the concept of a motor dealer as we know it today ceases to be relevant. Instead, manufacturers will appeal to fleet owners. Or will they?
Maybe taxi firms and other ride sharing business also take a hit, because manufacturers might cut out the middleman altogether and just operate their own vehicles as fleets. Maybe they won’t make any money from selling cars, and just from hiring them out and a journey-by-journey basis? Then, their production lines never produce cars they can’t sell, and only produce cars they need to refresh their own fleets. They compete with one another on the basis of the price of a journey, the entertainment or business facilities in the car, the comfort of the seating, and so on.
While we’re on the subject of journeys, there could be impact on motorway services and car park operators. Fewer stops may be required for tired drivers. What about refuelling (recharging)? A nationwide autonomous fleet operator might solve the problem of the passenger having to get involved in a recharging effort by simply having them change to another vehicle from their fleet that meets them en route, whilst the almost-depleted vehicle goes off to a fleet operator’s local centre to be recharged by a small team of people.
Why will people need to park cars at all, if all they have to do is call for one from a nearby autonomous fleet. Out of town parking in large areas of block-parked vehicles will become the norm, but what do people do with the things they need to take with them, like luggage, golf clubs, and so on. My car boot is always full of stuff (mainly belonging to the kids) that I don’t want to have to take out every time I send the autonomous car I don’t own off to graze in a field somewhere.
As I say, I’m fascinated by this stuff. It is far reaching, and about so much more than the readiness of the technology. It has massive implications for car manufacturing, liability assessment, insurance, ownership and the well-being of peripheral automotive and non-automotive businesses.