Debate House Prices
In order to help keep the Forum a useful, safe and friendly place for our users, discussions around non MoneySaving matters are no longer permitted. This includes wider debates about general house prices, the economy and politics. As a result, we have taken the decision to keep this board permanently closed, but it remains viewable for users who may find some useful information in it. Thank you for your understanding.
We'd like to remind Forumites to please avoid political debate on the Forum. This is to keep it a safe and useful space for MoneySaving discussions. Threads that are - or become - political in nature may be removed in line with the Forum’s rules. Thank you for your understanding.
Fully automated vehicles - 'not in our lifetime'?
HAMISH_MCTAVISH
Posts: 28,592 Forumite
I had a really interesting conversation (entirely by chance) recently with quite a senior manager at a car manufacturer when I happened to be sat beside him on a flight.
I asked him his thoughts on self-driving cars. Unsurprisingly he thought they were great and the way of the future.
I asked him how long he thought it would be before we were all sleeping on our way to work while the car did all the driving - or able to call our car to take us home autonomously after a night out drinking in the pub - rather than just the current party tricks of 'assisted driving' and 'auto parking'.
He said - and I quote - "Legally? Not in my lifetime".
The problem, as he explained it, is it not the technology... It's the liability and the politics around morality.
This chap reckoned a fully autonomous self-driving car could, in theory, be ready for market in the next 10 - 15 years. Some manufacturers claim to be able now, but according to him (and as the spate of recent accidents appear to show), they're not really and the technology is not yet ready.
But in his opinion it's no more than 10 or 15 years away from the technology being capable of fully autonomous driving more or less anywhere, and only 3-5 years away from it being capable of limited end-to-end autonomy (in the easiest environments for the technology to handle) but that would feel like proper self-driving in those circumstances.
However he then went on to explain the real problem - and this was the bit I found most interesting.
He said that no matter how good the technology, and no matter how good the programming, during the multiple decades where automated cars share the roads with non-automated vehicles or people - at some point an automated car will be forced to make a decision that has no good outcomes.
The choice - to simplify - is in the worst case scenario of an impending accident where the car has to choose between one of two bad outcomes - does the car make a decision that will likely kill it's driver and passengers... or does it choose to kill other road users instead.
If you programme it to protect the driver at all costs even if that kills other road users the political uproar and liability exposure would be unsurvivable for the manufacturer.
If you programme it to protect and avoid other road users at all costs even if it kills the driver then nobody would ever buy it - after all who will buy a car that you know will intentionally drive you off a cliff or into a tree if it thinks that's the only way to protect a cyclist?
And if you programme it to drive so slowly and defensively that it never risks killing anyone then traffic grinds to a halt everywhere and nobody will tolerate that either.
He reckoned there were only very limited solutions to the problem:
1. Every car on the road has to be fully automated and linked with each other using the same systems to predict conflicts and avoid any possibility of an accident from ever happening. And a national infrastructure spend of unimaginable proportions would have to be done to redesign all roads so as to minimise non-vehicle interactions.
2. The driver has to remain in ultimate control of the vehicle - making the same moral choices and carrying the same liability for those choices as we do today. Meaning no truly self-driving vehicles on our roads, just 'driver assist' technology, that will make things easier and safer but will not in any meaningful way reduce the number of drivers needed to drive cars in the UK.
His opinion was we'll eventually get to option 1 - but that's still 30 to 40 years or more away - and until then it'll be option 2 with some very limited exceptions for slow driving transport in segregated bus style lanes, etc, or possibly some automated HGV trailers in special motorway lanes as they're a more controlled environment.
To be fair I always thought the proper self-driving car thing was still a decade or two away - but I didn't realise that was likely optimistic and it could be more like half a century.:eek:
I asked him his thoughts on self-driving cars. Unsurprisingly he thought they were great and the way of the future.
I asked him how long he thought it would be before we were all sleeping on our way to work while the car did all the driving - or able to call our car to take us home autonomously after a night out drinking in the pub - rather than just the current party tricks of 'assisted driving' and 'auto parking'.
He said - and I quote - "Legally? Not in my lifetime".
The problem, as he explained it, is it not the technology... It's the liability and the politics around morality.
This chap reckoned a fully autonomous self-driving car could, in theory, be ready for market in the next 10 - 15 years. Some manufacturers claim to be able now, but according to him (and as the spate of recent accidents appear to show), they're not really and the technology is not yet ready.
But in his opinion it's no more than 10 or 15 years away from the technology being capable of fully autonomous driving more or less anywhere, and only 3-5 years away from it being capable of limited end-to-end autonomy (in the easiest environments for the technology to handle) but that would feel like proper self-driving in those circumstances.
However he then went on to explain the real problem - and this was the bit I found most interesting.
He said that no matter how good the technology, and no matter how good the programming, during the multiple decades where automated cars share the roads with non-automated vehicles or people - at some point an automated car will be forced to make a decision that has no good outcomes.
The choice - to simplify - is in the worst case scenario of an impending accident where the car has to choose between one of two bad outcomes - does the car make a decision that will likely kill it's driver and passengers... or does it choose to kill other road users instead.
If you programme it to protect the driver at all costs even if that kills other road users the political uproar and liability exposure would be unsurvivable for the manufacturer.
If you programme it to protect and avoid other road users at all costs even if it kills the driver then nobody would ever buy it - after all who will buy a car that you know will intentionally drive you off a cliff or into a tree if it thinks that's the only way to protect a cyclist?
And if you programme it to drive so slowly and defensively that it never risks killing anyone then traffic grinds to a halt everywhere and nobody will tolerate that either.
He reckoned there were only very limited solutions to the problem:
1. Every car on the road has to be fully automated and linked with each other using the same systems to predict conflicts and avoid any possibility of an accident from ever happening. And a national infrastructure spend of unimaginable proportions would have to be done to redesign all roads so as to minimise non-vehicle interactions.
2. The driver has to remain in ultimate control of the vehicle - making the same moral choices and carrying the same liability for those choices as we do today. Meaning no truly self-driving vehicles on our roads, just 'driver assist' technology, that will make things easier and safer but will not in any meaningful way reduce the number of drivers needed to drive cars in the UK.
His opinion was we'll eventually get to option 1 - but that's still 30 to 40 years or more away - and until then it'll be option 2 with some very limited exceptions for slow driving transport in segregated bus style lanes, etc, or possibly some automated HGV trailers in special motorway lanes as they're a more controlled environment.
To be fair I always thought the proper self-driving car thing was still a decade or two away - but I didn't realise that was likely optimistic and it could be more like half a century.:eek:
“The great enemy of the truth is very often not the lie – deliberate, contrived, and dishonest – but the myth, persistent, persuasive, and unrealistic.
Belief in myths allows the comfort of opinion without the discomfort of thought.”
-- President John F. Kennedy”
Belief in myths allows the comfort of opinion without the discomfort of thought.”
-- President John F. Kennedy”
0
Comments
-
Regulators always get a bad name but they are not stupid and self destructive.
A fleet owner just needs to show the overall rate of accidents is lower than human drivers.
Also insurance exists for everything
It doesn't matter if a car kills its passengers or the public because Human drivers do the same
Its not such a common and such a horrific thing so as to stop human drovers from driving.
The cost of these accidents is about 5p a mile (£500 car insurance over 10,000 miles)
Its just that a cost and self drive cars will also have accident costs but likely closer to 1p a mile
There are also potentially huge side benefits that the governments and public will see like much less need for parking and car infrastructure. Law enforcement. Traffic management. Etc
My guess is they will be very common by 2030 and take close to full market share by 20400 -
HAMISH_MCTAVISH wrote: »The choice - to simplify - is in the worst case scenario of an impending accident where the car has to choose between one of two bad outcomes - does the car make a decision that will likely kill it's driver and passengers... or does it choose to kill other road users instead.
I've been banging about this for years. What would likely happen is that different versions of the software would be available depending who you were. If you were a nobody you'd get the version that killed you whereas if you were Tony Blair you'd get the version that protected you at all costs.
It's utopian and won't happen.0 -
There's certainly the option that diplomatic self driving cars might have different rules, but it's the unrealistic expectations that'll kill it off - people don't want to trust machines unless they are flawless, rather than just being less flawed than people.0
-
There's certainly the option that diplomatic self driving cars might have different rules, but it's the unrealistic expectations that'll kill it off - people don't want to trust machines unless they are flawless, rather than just being less flawed than people.
Indeed.
I can absolutely see a form of driver assist mode where you get in the driving seat, enter a destination, and the car will drive you there in the next decade or so.
But only if you, as a driver, remain legally in control of the vehicle and ready to take over should something untoward happen.
Truly autonomous vehicles, ones that [STRIKE]can[/STRIKE] can legally drive you to work while you're sleeping or back home from the pub after a big night out, may well not happen in my lifetime.
Another example of technological dreams being defeated by cold hard reality.
On a positive note however the young taxi drivers and driving instructors of today can look forward to a long career still ahead of them.
It seems the robots aren't coming to take their jobs... For another half century or so anyway.;)“The great enemy of the truth is very often not the lie – deliberate, contrived, and dishonest – but the myth, persistent, persuasive, and unrealistic.
Belief in myths allows the comfort of opinion without the discomfort of thought.”
-- President John F. Kennedy”0 -
Not sure if it will happen in our lifetime, but I surely hope it doesn't.
There's a certain charm to revving the V8 while aiming a well directed groin-thrust or fist-pump at some unsuspecting renter.
Don't blame me, I voted Remain.0 -
We already have the beginnings of AI in fifty years time humans might not even be biology anymore.0
-
It all comes down to price
Do you go for the human uber at £1 a mile or the robot uber at 20p a mile?
Moreover the network effects mean not only price but convenience. Do you wait 10 minutes for the human uber driver or 1 min for the robot uber?0 -
ilovehouses wrote: »Who says you'll get the choice? The London Underground could be automated fairly easily and be run at a lower cost but I bet you'll still have drivers and guards etc. for years to come - for your safety of course..
Exactly.
Aviation is decades ahead here with automation and arguably has a much easier task to do than self driving cars have - no pedestrians in the sky, just take off, fly to destination, and land, with very few objects up there to hit - but we still have pilots....
In fact we have a global shortage of pilots.“The great enemy of the truth is very often not the lie – deliberate, contrived, and dishonest – but the myth, persistent, persuasive, and unrealistic.
Belief in myths allows the comfort of opinion without the discomfort of thought.”
-- President John F. Kennedy”0 -
I went on the DLR last week.0
-
It all comes down to price
Do you go for the human uber at £1 a mile or the robot uber at 20p a mile?
Moreover the network effects mean not only price but convenience. Do you wait 10 minutes for the human uber driver or 1 min for the robot uber?
Indeed.
Autonomous vehicles offer the opportunity to change the ownership and usage model.
The country that masters the self drive road network can sell that expertise to any of the major cities around the world.0
This discussion has been closed.
Confirm your email address to Create Threads and Reply
Categories
- All Categories
- 347.8K Banking & Borrowing
- 251.9K Reduce Debt & Boost Income
- 452.2K Spending & Discounts
- 240.1K Work, Benefits & Business
- 616.3K Mortgages, Homes & Bills
- 175.4K Life & Family
- 253.5K Travel & Transport
- 1.5M Hobbies & Leisure
- 16K Discuss & Feedback
- 15.1K Coronavirus Support Boards