http://www.techradar.com/news/car-tech/ ... od-1275805

I want them here and I want them now. What I don't want is to take a test

I expect that if they did allow privately owned cars, you'd need to be a proper driver so you can take over if necessary. Who, in an accident, is liable? the owner, the passenger, the software company, the manufacturer? No one has answer that one yet.

The question of liability in the event of an accident is indeed a very interesting one.
If someone can take over the vehicles operation then it's not a driverless vehicle surely, just one with very advanced driver assistance.
A true driverless vehicle shouldn't really have any way for the occupants to take over the vehicle.

Depends if you have to be at the controls or not. There's an obvious reaction difference between

a) having your hands on the wheel and paying attention then realising you have to intervene
b) looking up from your smartphone in time to realise you have to take control and then grabbing the wheel

The former, fine, but in that case why bother with the driverless car? if we're all sat in front of the wheel attending to the traffic on the off chance that something bad might happen, the benefit of the driverless car seems pretty minimal to me.

In the latter case there are plenty of possible accidents where there's no way the passenger could intervene in time.


Well, at one end of the scale, yes. There's a whole continuum of 'driverless' from basically a self driving box you sit in to a pretty normal car with extensive driver aids to massively decrease the accident risk. The latter isn't what most people would consider 'driverless' though.

I would consider the former driverless. The later is driver assist.

The only way driverless cars can work 100% is when ALL cars are driverless. If all cars are driverless and can communicate between each other then it will pretty much remove any and all accidents. The only possible accidents then would be pedestrians jumping in the way of cars. And still they'd have a better chance than if a person was driving.

You could also have cars driving faster and closer together. The main delay in stopping a car is the reaction of the person behind the wheel.

If the car 10 in front can say there's been a crash then the car can slow down before a person would even be able to see it.

Merging would be a dream. Seamlessly linking together like a zip. Etc...

I don’t see why all cars would need to be driverless – sure if they were you would have advantages but no need for them all to be

The Driverless Car will still need to be able to react to events – e.g. people stepping out in front of them / cyclists etc

You could however make the owning of the driverless car more “desirable” via cheaper insurance / Road Tax, higher speed limits etc

I think Foggy's point was that unless all cars are driverless, you're going to get legal issues of responsibility in accidents. His logic being once all cars are driverless and we've taken the fallible fleshy bits entirely out of the equation, the number of accidents will become very small and we can just have some level of assigned responsibility for the few that happen.


I don't know about higher speed limits - we don't want to offset the reduced number of accidents we have by making the ones we do have more lethal - but once we have enough statistics on the rate of accidents of driverless vs driven vehicles, we could certainly modify insurance premiums appropriately.

Road tax is primarily based upon emissions so a driverless car of a given engine size would pay the same tax as a driven one, unless we change the system somehow. It seems likely to me though that by the time driverless cars are actually a commercial reality, the majority of driverless cars will either be electric, fuel cell or have small, highly efficient petrol engines and therefore pay little road tax anyway. Nobody's going to build a driverless Ferrari, there's utterly no point.