The Coming Robot Apocalypse of Car Insurers

wymondham

purch wrote: »

I never read threads with the word Apocalypse in the title :eek:

evidence suggests differently?

zagubov

Generali wrote: »

http://ftalphaville.ft.com/2014/03/04/1787962/insurers-will-destroy-themselves-to-nudge-us-into-robot-utopia/

The simple version is that insurers will give cheaper insurance to cars that have automated features to stop them crashing. In doing so they will make 'crash proof' cars much more attractive.

Crash proof cars don't need expensive insurance and the likelihood is that we will not own cars but instead borrow them from robot manufacturers like Roomba and Google. These companies will need insurance, not drivers. As a result Britain will go from having perhaps 40,000,000 insurance policies to having perhaps 20.

Not great if you work for a retail insurance company.

Retail insurance can flourish in the most barren environments. Years ago in the Soviet Union it was one of the few self-employed businesses you could do privately (although there was ultimately one giant insurance comapny behind it). People would go door to door and sell people insurance for their camel, donkey, car, white goods, whatever (those examples are for real by the way).

People hadn't many goods to spend their money on so the state generated financial products to spend their money on instead, and the policies protected the precious items and paid for their repair, which, due to poor quality, in many cases was frequently required.

I reckon they'll come up with new insurance products for our virtual possessions or things we haven't even thought of yet.

bugslet

ruggedtoast wrote: »

This all looks great for googles R&D I'm sure but unless they ban human drivers from the road in one go I think we are decades away from any kind of real self driving car.

I agree with the above, but I think somewhere along the line it will happen.

http://www.youtube.com/watch?v=oV3nINN2ELQ

In addition to the above, AIUI, Volvo have got the go ahead to trial 10 trucks on a public highway this year, which will be 'driven' by one driver in the first vehicle that will then communicate with the following 9.

I presume they will shut the motorway!

System

ruggedtoast wrote: »

I'm curious as to how this technology can cope with unpredictable other human drivers.

I'm wondering particularly about situations in crowded city traffic where you can't pull out unless you get a nod or a flash from another driver.

The non-robot driver will simply pull out anyway, knowing that the robot will react in a non-human, non-aggressive way, and give way.
Robots will drive so perfectly there will be no need for minority human drivers to take any care or follow any rules.

Generali

Clifford_Pope wrote: »

The non-robot driver will simply pull out anyway, knowing that the robot will react in a non-human, non-aggressive way, and give way.
Robots will drive so perfectly there will be no need for minority human drivers to take any care or follow any rules.

Indeed.

And that will mean a higher proportion of collisions will involve non automated cars which means higher premia for drivers which forces drivers into automation.

JWF

They will have to work pretty hard to prevent any 'butterfly effect' interactions, where the changing of lanes by one car results in a huge pile up slightly further on/later. In theory it's all very predicatable, safe and boring. In reality it is an immensely complex system with all sorts of feedback going on.

If I thought it was even feasible, I'd agree that it will probably bump up insurance for human drivers once the robotic drivers reach a certain concentration and effectively dictate how traffic flows & vehicles interact.

Flipping the bird at a robot won't have quite the same calming effect as it does to a human driver!

Conrad

Could a computer tell a sheep from a small fog patch in a dip? Would it slam on the brakes as it approached and thus cause the car behind to smash into my rear?


Would it notice a big board of plywood about to blow off that lorry 50 metres in front of you on the motorway?

cells

Generali wrote: »

http://ftalphaville.ft.com/2014/03/04/1787962/insurers-will-destroy-themselves-to-nudge-us-into-robot-utopia/

The simple version is that insurers will give cheaper insurance to cars that have automated features to stop them crashing. In doing so they will make 'crash proof' cars much more attractive.

Crash proof cars don't need expensive insurance and the likelihood is that we will not own cars but instead borrow them from robot manufacturers like Roomba and Google. These companies will need insurance, not drivers. As a result Britain will go from having perhaps 40,000,000 insurance policies to having perhaps 20.

Not great if you work for a retail insurance company.

Yep self drive vehicles will reduce the nation's insurance burden from its current circa £15B a year to maybe just £1B a year. A big part of the savings will simple be less accidents resulting in much lower claims but also a lot lower overheads in not needed to advertise and communicate with 35million motorists but instead woth a handful of robocar makers.

There is also the fuel savings benefits. A typical human car uses about 200 gallons a year while a robo car will likely reduce this to under 100 gallons. Saving another £15B or so....

Plus the UK currently buys near 2million new cars a year to keep its stock at this level. This will likely more than half as the robocars will be utolised more efficiently. Again saving about £15B a year...

there are mutiple other benifits and aacings but these three alone are close to a 50B a year saving. A huge huge amount


it is a tech the world shoild push to make a reality ASAP

cells

ruggedtoast wrote: »

I'm curious as to how this technology can cope with unpredictable other human drivers.

I'm wondering particularly about situations in crowded city traffic where you can't pull out unless you get a nod or a flash from another driver.

With computer cars traffic would be drastically reduced

as for problems with computer cars. They will not be perfect but we don't need them to be perfect we just need them to be better than humans and that they almost certainly will be.

also computer cars can probably deploy active safety features to lower accidents further. For instance if a break failure happened in yur car it could send a signal to the other robo cars to clear the next half mile so the car can stop from air and road drag or even use another robo car to stop your one or just send a text to the ambulance services of the crash milliseconds before or after the incident

cells

Conrad wrote: »

Could a computer tell a sheep from a small fog patch in a dienough eit slam on the brakes as it approached and thus cause the car behind to smash into my rear?


Would it notice a big board of plywood about to blow off that lorry 50 metres in front of you on the motorway?

Most likely in time they will.

A computer car will likely be used on fuel saving mode. So unlike most humans if it detected a stationary object x meters in front it would let go of the gas so that it arrives to a stop very gently before the object. To minimise the break use and hence fuel usage so no harsh breaking. At some point of course it would realise that the fog is just fog assuming it didn't know in thr first place

a better one is if a person or animal jumps out into the road. First that is a problem for both humans and computers. The difference is a human panicks and slams the breaks an does little else. The computer could engage the breaks about half a second quicker but it could also calculate and try to sway the car to avoid the collision.

also the number of cars on the road will likely half or more which in itself woild massively reduce accidents.

Also the robo cars will likely have a 3D map of all the roads so they will know all the bimps himps and pot holes and drive to avoid them.



Also note an old engineering saying. Perfection is the enemy of the good.
what that means is you cannot build perfection so dont let that stop you from building the good. Self drive tech will not be perfect it will be good and better than human drivers