Electric cars.

[dad_of_jon]

There is also a significant moral issue around the decision making that must occur at the point the AI driving the vehicle realises it has to either sacrifice it's occupants or another human as an accident is unavoidable.

It's an interesting argument, I believe that the AI in control if correctly programmed would not allow itself to be driving in such a way.

Of course there is a statistical chance than a nun, a pregnant mother wheeling triplets in a pram attempt to cross a motorway on foot where the autonomous car who's sole occupant is a convicted murderer and drug dealer travelling at 70mph. (or usually 50mph in wales) has to make a decision.

We also may not be taking into account that the passive safety in vehicles is such that if obeying the traffic regulation no-one in the vehicle would be seriously harmed even if evasive maneuvers were required.

 

[Stu's Brews]

That decision would not be made by artificial intelligence sit would be programmed into the car's driving software by a a person following a police decision.
If you were driving would you sacrifice the car occupants or the other person? The probability an event seems rather remote, though.
If it were proven that autonomous driving reduced road fatalities overall, what would your company do?

The problem with the unavoidable accident situation is the legal implications. It's a conscious decision taken by the vehicle manufacturer when developing the AI. That means it can definitely be challenged in court and not many manufacturers would want to take their chances on that.

Regarding reducing road fatalities overall, that is something we are very much involved in developing AI to support. There are 5 acknowledged levels of autonomous driving with level 5 being fully autonomous but everything up to that being either accident avoidance or autonomy in controlled situations (such as the aforementioned motorways). We intend to develop up to level 4 as that will make a significant impact on accident avoidance.

 

[Stu's Brews]

It's an interesting argument, I believe that the AI in control if correctly programmed would not allow itself to be driving in such a way.

Of course there is a statistical chance than a nun, a pregnant mother wheeling triplets in a pram attempt to cross a motorway on foot where the autonomous car who's sole occupant is a convicted murderer and drug dealer travelling at 70mph. (or usually 50mph in wales) has to make a decision.

We also may not be taking into account that the passive safety in vehicles is such that if obeying the traffic regulation no-one in the vehicle would be seriously harmed even if evasive maneuvers were required.

When you dig into the detail you realise there are millions of potential situations the AI may have to deal with which makes it impossible to avoid accidents.

Everything from the weather, road signage, roadworks, unexpected human actions, etc. can all cause havoc as there are so many variations of the same thing. Which is why keeping a human involved to some level is necessary.

Passive safety has pretty much reached the limit of what can be realistically achieved as we're now in a situation where adding more mass to vehicles simply increases the energy involved in an accident. So now the role is for active safety to take over and help avoid the accident in the first place.

 

[Chippy_Tea]

 

[Stu's Brews]

https://www-autoblog-com.cdn.amppro...8/autonomous-cars-slow-progress-losses-doubt/
https://www.wired.com/story/ford-abandons-the-self-driving-road-to-nowhere/
A couple of good articles with details on the latest thinking for autonomous driving.

 

[Bernie]

https://www-autoblog-com.cdn.amppro...8/autonomous-cars-slow-progress-losses-doubt/
https://www.wired.com/story/ford-abandons-the-self-driving-road-to-nowhere/
A couple of good articles with details on the latest thinking for autonomous driving.

I haven't read the articles but defining bona fide research into autonomous driving as a scam, whatever the cost, takes me back to the point I made earlier that car drivers are reactionary and are hostile to any change that would disturb their opinion of what driving should be.

 

[Stu's Brews]

I haven't read the articles but defining bona fide research into autonomous driving as a scam, whatever the cost, takes me back to the point I made earlier that car drivers are reactionary and are hostile to any change that would disturb their opinion of what driving should be.

Yeah the headline is rather sensationalist but the contents of the article are very good. It focuses on some of the issues encountered so far with the technology and how a previously very eminent engineer in the field of has now switched to a position of doubting the feasibility of autonomous driving.

 

[Chippy_Tea]

how a previously very eminent engineer in the field of has now switched to a position of doubting the feasibility autonomous driving

For me it's not just about the safety of driverless vehicles which i do find a worry it's about jobs, when the 20 million jobs predicted below are being done by robots what are those that used to do them going to do to earn a living.


Robots could take over 20 million manufacturing jobs around the world by 2030, economists claimed Wednesday.
According to a new study from Oxford Economics, within the next 11 years there could be 14 million robots put to work in China alone.
Economists analyzed long-term trends around the uptake of automation in the workplace, noting that the number of robots in use worldwide increased threefold over the past two decades to 2.25 million.
https://www.cnbc.com/2019/06/26/robots-could-take-over-20-million-jobs-by-2030-study-claims.html

 

[Stu's Brews]

For me it's not just about the safety of driverless vehicles which i do find a worry it's about jobs, when the 20 million jobs predicted below are being done by robots what are those that used to do them going to do to earn a living.
Robots could take over 20 million manufacturing jobs around the world by 2030, economists claimed Wednesday.
According to a new study from Oxford Economics, within the next 11 years there could be 14 million robots put to work in China alone.
Economists analyzed long-term trends around the uptake of automation in the workplace, noting that the number of robots in use worldwide increased threefold over the past two decades to 2.25 million.
https://www.cnbc.com/2019/06/26/robots-could-take-over-20-million-jobs-by-2030-study-claims.html

Completely agree with you on that. As a business owner it makes complete sense to use robots that are effectively free instead of salaried employees.

However, is a significant human cost to that and it will hit the lowest earners first as they generally the ones doing tasks that could be automated.

Potentially it could help reduce the cost of products but I fear the reality would be that it would instead boost profits.

 

[Rodcx500z]

And why would you want to own a car that drives itself, how boring is that, just sat there watching a movie or doing a crossword naw not for me

 

[Chippy_Tea]

And why would you want to own a car that drives itself, how boring is that, just sat there watching a movie or doing a crossword naw not for me

Spot on i enjoy driving, if we are going to go driverless we may as well all have a small car with the same spec at least it would put a stop to the one occupant Chelsea tractors and huge SUV's that seem to be becoming more and more popular today.

 

[Chippy_Tea]

 

[Rodcx500z]

,

 

[Leon103]

And why would you want to own a car that drives itself, how boring is that, just sat there watching a movie or doing a crossword naw not for me

Why own it. Cars will be owned by companies such as Uber and we would be passengers

 

[Rodcx500z]

Why own it. Cars will be owned by companies such as Uber and we would be passengers

Naw i will never be a passenger

 

[Bernie]

Completely agree with you on that. As a business owner it makes complete sense to use robots that are effectively free instead of salaried employees.

However, is a significant human cost to that and it will hit the lowest earners first as they generally the ones doing tasks that could be automated.

Potentially it could help reduce the cost of products but I fear the reality would be that it would instead boost profits.

It was ever so. All kinds of jobs are automated; a doctor will send you off to a machine for a blood test, lawyers can call up databases automatically to look up cases, aeroplanes are run on autopilot most of the time till human error is needed.
It does not take expert economists to see that all this can be traced back to early farmers who used dogs to herd animals and beasts of burden to pull ploughs thus making hunter-gatherers redundant.

 

[dad_of_jon]

And why would you want to own a car that drives itself, how boring is that, just sat there watching a movie or doing a crossword naw not for me

At some point most people become unsafe to drive on the roads and if the choice is a driverless car or a taxi which itself could be driverless - because the walk from your house to public transport is too far away what would you choose?

 

[Bernie]

At some point most people become unsafe to drive on the roads and if the choice is a driverless car or a taxi which itself could be driverless - because the walk from your house to public transport is too far away what would you choose?

Drivers between the ages of 17 and 25 are more than 6 times more likely to be involved in an accident. Driverless cars should be used for all age groups. Collisions involving older drivers increases markedly above 80 years of age.
Most collisions involve drivers under 50.



https://www.gov.uk/government/stati...es-great-britain-older-drivers-factsheet-2020

 

[Rodcx500z]

At some point most people become unsafe to drive on the roads and if the choice is a driverless car or a taxi which itself could be driverless - because the walk from your house to public transport is too far away what would you choose?

Hopefully uthanasia will be legal by then

 

[dad_of_jon]

When you dig into the detail you realise there are millions of potential situations the AI may have to deal with which makes it impossible to avoid accidents.

Everything from the weather, road signage, roadworks, unexpected human actions, etc. can all cause havoc as there are so many variations of the same thing. Which is why keeping a human involved to some level is necessary.

Passive safety has pretty much reached the limit of what can be realistically achieved as we're now in a situation where adding more mass to vehicles simply increases the energy involved in an accident. So now the role is for active safety to take over and help avoid the accident in the first place.

of course , 2 tonnes travelling at any given speed is going to contain more kinetic energy than a 1 tonne vehicle.

maybe in the future there will be a competence driving test where the candidate will undertake a simulated driving test at the same time as AI. if you score better than the AI you get to drive. If not then the AI does.

My vitara has lane guidance and it will literally drive you around a bend - IF the road edges are marked with white paint. It only gets confused when one lane opens up into 2 so the road width doubles before the lane is marked out for left turn and straight-on/right. It tries to steer you down the middle. So you need a better system. If it had been integrated with the sat nav it would also know that I usually take a left 99% of the time so just position for the left lane.

I have appeared to drift off topic so, I'm hoping not to buy another car whatever powers it. I think I'd rather put a solar roof on first when the roof needs redoing.