After yesterday’s post I got thinking more about our “driverless” future (ironically, while on a long drive in my 16-year-old Nissan).

Truth is, there is no “driverless” future: there’s still a driver, we just have to think differently about who it is.

Imagine a future scenario: I get into the car, type a postcode into the satnav, the car sets off an I fall asleep in the back, which in a future of fully autonomous cars, I’ll be able to do, right?

A crash happens. Maybe my car drives into someone, maybe vice versa. (Some predict autonomous vehicles will usher in a future without crashes – but in the intervening years, you’ll have human drivers hitting driverless cars). All I know is that I wake up surrounded by broken glass.

After the dust settles, I send in a claim to my insurer (with whom I have fully comprehensive cover). Right now, in 2017, the insurer will investigate the accident to see if they can blame me for it, in which case they might be able to avoid having to pay (or at the very least, they put up my premium). Or blame the other driver and heap the cost onto them.

But in a “driverless” future, the insurer can’t turn down my claim, because I had nothing to do with the accident. I was literally asleep at the time. Instead of investigating the humans, the insurer has to start looking at the car’s autonomous behaviour. So for a start, they’re gonna have to start hiring software analysts to interpret the data. And then they’ll have to answer a key question: did the car behave within the boundaries set by the code? (In which case, the code was bad to start with) Or did the car somehow go outside the code? (In which case, how?)

In order to do this, ideally the insurer will need access to the car’s software source code, which is of course owned by the manufacturer. Is the car maker going to hand it over? Seems unlikely; partly because it’s the manufacturer’s intellectual property, and partly because of what the insurer will be wanting to do next….

To escape liability, the insurer will want to blame the manufacturer (“it was your bad code/bad implementation that caused the crash”). This is a big switch for the car industry: at the moment, as a manufacturer, you simply sell the car and that’s it, deal done. How will car makers feel about a future where they’re perpetually on the hook for any crash their product is involved in? (We’ve already seen how that played out for Tesla).

At a bare minimum, car makers will insist on regular code updates for the car. So will that be included in the initial price, or will we eventually have to pay for a digital MOT each year?

Part of the shift here is about the ownership-liability model. At the moment, because I own my Nissan, I’m also liable for driving it, and so I’m liable for insuring it. In the future I will still own the car, but effectively I won’t be liable for driving it (this throws up some interesting edge cases: for example, I will probably still be liable for pumping up the tyres. So if I fail to do so, perhaps the car will refuse to drive, because my digital chauffeur is effectively refusing to take liability for driving on my poorly maintained wheels….).

But if I’m not liable for driving the car, why should I be liable for insuring it?

There’s possible way out of all this mind-boggling stuff: change the ownership model. I suspect we’re heading for a future in which people don’t own cars, but simply lease them in some way. (This is part of a technology mega-trend: the shift from owning to renting, which I’ve blogged about before)

In a way, shared ownership of cars makes a lot of sense, especially given some estimates that cars are parked 95% of their lives

But there’s a couple of issues with the idea of a leasehold future for cars. Firstly, I’ve been a property tenant most of my adult life, and it’s always been a pretty dis-empowering experience, with landlords failing to maintain and improve houses, putting up rents, and bringing tenancies of much-loved properties to an end (all this, despite having paid as much in rent as I now do in mortgage).

But there’s a deeper ramification for the car industry: how to make and sell cars for a rental market.

At the moment, car building and marketing emphasises ownership values (cool bodywork, fun gadgets, leather seats, etc). How much of this will apply in a rental market? When you order an Uber, do you really care what model of car it is? Sure, for longer trips you might want to upgrade, but for most trips, you’re looking for price and convenience rather than a swish car.

Will a leasehold car future mean vehicles made more for reliability and share-ability? And if you’re still mulling how this will all work, here’s a big question about the future of shared cars: who’s going to clean them?