Driverless Cars- Scariest Thing Ever?

[Noofly]

Am I the only one completely freaked out by the thought of driverless cars? I can't even bear it when DH uses the self parking gadget on his car. How am I going to cope in a world of robot cars?

(lighthearted in that yes, I will obviously cope, but am I alone in thinking Eeek?!?!?)

Computers can go wrong (anyone here ever had a computer that never crashed?)

Computers have no common sense (Satnav trying to direct people over cliffs, down one way streets etc).

There are many unexpected things that could happen, from roadblocks to snowdrifts. Just one example - if a set of traffic lights failed, the driverless car wouldn't be able to follow hand signals from a police officer.

Computers are at risk of being hacked, and troublemakers could interefere with the program to cause dangerous things to happen. Perhaps viruses and malware could also be a problem.

Loss of privacy, as your car may be in contact with data centres. Your location could be found by someone who misuses, or hacks into, the system.

In the event of a glitch, drivers may be out of practice, leading to poorer standards of driving on the roads from human drivers.

Power cuts happen.

They're made by Google.

The driverless car is an interesting idea and I like the sound of relaxing instead of having to drive. But I just don't trust them yet. I will take the train instead of a driverless car.

There are many unexpected things that could happen, from roadblocks to snowdrifts. Just one example - if a set of traffic lights failed, the driverless car wouldn't be able to follow hand signals from a police officer

As they will be interconnected, one car that flags up a roadblock will make other cars re calculate and take another route, once we have driverless cars we will not need traffic lights.

Comparing cars to computers is not in the same ball park, as for tracking that already happens (and will be more widespread if the snoopers charter goes through) via your mobile phone. Data is already collected as to where you are either by just having location services switched on, or by using your mobile data or hooking up to your nearest free wifi.

Google is the one making them now has they have more money than know what to do with and a lot of technology resources at hand.

Power cuts happen and yet they don't bring down data centres, so another mute point.

justanother around 80% of the cars in this country are imported and about 80% of cars built in this country are exported.
All cars built for the UK can only be sold in the UK as the steering wheel is on the wrong side. There is no first mover advantage for UK companies.
What is the 'least worst option'? A pensioner driver is killed to save a toddler, but a working age driver is protected and the child is killed?
At what point does the car I own decide my life is less important than someone elses?

There is no first mover advantage for UK companies

Of course there is a first mover advantage, you think the UK manufacturing process can only build cars for cars that drive on the correct side?

I'm not saying the uk will gain the manufacturing etc, but there is obviously a first mover advantage for whoever takes this up, if we are the first to fully legalise then we are in a better position than not, it may not be for direct manufacture but it could be any other accociated service, real world R&D or just infrastructure redesign.

As for least worst it's the age old trolley argument that I posted up thread and it's one of the first things students learn when moving into AI, machine learning has moved on leaps and bounds.

There is no right answer, what would you do, if you was in that position?

Most humans either panic and kill everyone or self preservation kicks in, if it was a case of avoiding a child and moving into oncoming trucks and buses or going the other way into parked cars then the computer would make the correct choice every time whereas humans don't.

In the future there will be no need for steering wheels which will make the inside of the car safer and with cars communicating with each other there could be an element of forewarning going on, the fact that you can garuntee that all the cars will be doing exactly the speed limit is another safety valve.

once we have driverless cars we will not need traffic lights.

How will pedestrians get across the road?

Power cuts happen and yet they don't bring down data centres

Neah, that would be Denial of Service attacks bringing down websites.... I hear Rackspace (H U G E data service) went down due to a truck driving into the building. Fire fighters insisted everything be turned off, even the back up power supply.

I love driving and will never have a driverless car.

I have also seen Terminator. Driverless cars are a little too Skynet for me.

I love the idea of driverless cars. I think in the long run they will be safer. If you take a step back and think about millions of cars driven by people who are tired, drunk, angry or distracted I'd say THAT was scary.

However right now there are some things a computer can't do as well as humans. Some of those won't apply often on a motorway, but would in a residential area and especially if there are cars parked both sides. (it must actually be easier to design a plane autopilot)

Your eyes/brain can take in a lot of information, identify people and objects and predict likely actions. A computer can detect an obstacle, but probably can't (yet) tell the difference between a tesco carrier bag blowing in the wind, a toddler and a fox. No matter how good the programmers are someone has to decide what should be done in such circumstances. Do you instruct the car to assume it's a toddler every time or rely on 'best guess'?

A human driver can see that a child on the pavement is running towards the road and slow down in case the child doesn't stop. You can't expect the car to make judgements like that.

On the other hand maybe the mistakes it makes (that might have been avoided with a human driver) will be still be much fewer then those made now by humans. Could we accept that and call it an improvement? It's logical if a little uncomfortable and the system will improve over time while people won't.

Your eyes/brain can take in a lot of information, identify people and objects and predict likely actions. A computer can detect an obstacle, but probably can't (yet) tell the difference between a tesco carrier bag blowing in the wind, a toddler and a fox. No matter how good the programmers are someone has to decide what should be done in such circumstances. Do you instruct the car to assume it's a toddler every time or rely on 'best guess'?

It's a nice theory, but your wrong, of course the car can detect the difference between a bag and a child or fox, that was true maybe 10 years ago but not now, the care also has better vision than you what with multiple camera angles and a 360 swoop all being procesed a lot quicker than you can scan your eyes over an area. With over 1 million miles done and only one accident where the car was at fault and that could be said was due the bullying nature of the bus driver, I think a computers has come on a bit since you last took a look.

Neah, that would be Denial of Service attacks bringing down websites.... I hear Rackspace (H U G E data service) went down due to a truck driving into the building. Fire fighters insisted everything be turned off, even the back up power supply.

Ddos attacks against websites are not in anyway compatible, you could try and attack a distrubtuted data centre but we are talking China level attacks.

I visit rack space regularly and even though there was an outage DR had already taken hold and the critical system where not affected.

DH's company website with Rackspace went down for 12 hours. They were a big client at the time, income stream from click thrus, so not best chuffed.

Actually sorry whistle down, I re read my reply it seems I'm being a bit harsh, not intentional sorry.

I may have crap English and bad spelling but I'm not having a go.

You also have to consider that it's not some software devs job to almost show it pictures of bags floating in the wind and a child or fox and every possibility in between, machine learning comes into it heavily, which doesn't really illeviate many fears on this thread but it does know the difference between objects.

Again, comparing websites with critical systems is chalk and cheese.

Web hosting and government/safety specific systems are not co located, parts may be distributed to these data centres, but it will be one of many entry points so I fail to see what your trying to say.

No one cares when the web front end goes down (in real terms) but attacking some infrastructure such as the driverless cars network would be a military style attack and if that happens worrying about your car is the least thing to worry about in the grand scheme of things (unless your in it perhaps)

In the case of rack space DR had already taken hold for critical systems, and since this outage we have mirrored data centres that would compensate driving a truck into it.

Well cars already have active cruise control....they lock onto the car in front and set a certain distance behind. They will auto break if anything comes into sensor view...sometimes a little disconcerting if you are going round a sharp bend with cars parked..

justanother computer would make the correct choice

There is no 'correct' choice in a moral dilemma. There is the cheapest choice (normally killing the child), a 'societal choice - such as killing a pensioner over a working age person, and other choices.

A computer may be better able to make a choice than a human in a split second, but who decides what the 'correct' choice is? the owner, the manufacturer, the insurer, the government?

only one accident where the car was at fault and that could be said was due the bullying nature of the bus driver - so driverless cars are ok, as long as there are no drivers sharing the road with them?

Hasn't anyone watched Terminator??!!

They will all become self-aware and unleash the nuclear apocalypse and continue to hunt us to extinction!!

I said there was no correct choice in a moral dilemma carol?

However in the scenario I suggested where a child ran out into the road and in order to save that child the solution is to drive into oncoming traffic where a bus or lorry is travelling and injuring more people or causing a bigger situation, or go into the parked car possibly injuring yourself the car would make IMO the morally right choice and go into the car 100% of the time, humans probably 50/50.

When a car accident happens the first thing to run through someone's mind is not of a societal nature obviously, after the initial jolt it's self preservation. If my gran is in a driverless car and some toddler runs out in the road, there are no algorithms that go 'well the old dear has had a good life so fuck it', it will take what it has machined learned to be the least worst outcome in that scenario.

It's not going to alleviate your fears of putting yourself in control of something else, but the system will make a calculated decision at that point in time from what it or other cars has learned, there is no possible algorithm that could foresee every conceivable scenario. And if there was we would take years for the public to decide which was the correct course of action to take on any given scenario. However the car will take an actual calculated decision not just for you but for the objects around you, and will do so quicker than a human ever could.

As for the million miles comment, I wasn't suggesting that at all.

It was in my reply to the poster that suggested people can scan built up areas better than a computer ever could, as the computer would not be able to differentiate between a serious potential accident or just be startled by a bag blowing across the road. The cars have done over a million miles in testing, many thousands in built up areas and around heavy traffic, of all the accidents that the cars has been involved in(which is very small considering) only one was due to the car, and that could be said to be of bullying nature from the bus driver.

They will all become self-aware and unleash the nuclear apocalypse and continue to hunt us to extinction!!

Until they can transform I'm alright as I have steps to my front door

Yep MrsFlorrick i said that at the start of the thread. Terrifying.

the govt likes the idea of driverless cars because it will make it easier for them to know the details of every single journey that car takes - and once they know those details it will a small step towards controlling what journeys that car is allowed to make.

the govt likes the idea of driverless cars because it will make it easier for them to know the details of every single journey that car takes - and once they know those details it will a small step towards controlling what journeys that car is allowed to make

Oh come on, that is tin foil hat stuff, people are a lot more active in giving out geo tagging data, from Facebook,Instagram etc through to Norwich Union saying they will give you cheaper insurance, that's excluding the dragnet that's being employed by gchq/NSA

I have seen minimal threads on here about the snoopers charter and those that have appeared, the prevailing consusus is that "if you do nothing wrong, you've nothing to worry about" driverless cars may report back but let's not try and pretend this is some nefarious government push considering what is available to them already.

This is a technological advancement first and foremost, theres already enough data being collected that this will just pinpoint your route to location b more acuretly and what speed you was going. It won't prevent you getting on a bike or getting to your location on foot...

I like the idea of driverless cars. I was driving yesterday when a man overtook me and then pulled in so I have to brake sharply. When I pulled out to overtake him back, I saw he was peeling an orange and on the phone while driving at 70.

Driving won't be banned imo. It will become an uninsurable hobby like deep sea diving and off piste skiing. You can choose to drive yourself bug you'll need to be a billionaire to pay the premiums.

AFAIK, the current thinking on 'driverless' cars that we're going to see in the short term, is going to require a legal driver behind the wheel much as with pilots. They can take back control at any time.
That's likely to be the scenario for the next decade, minimum.
Cars can be hacked now. I can't remember which manufacturer it was that recently suspended it's interactive service because it was easily hacked.
Governments and insurers are going to require all sorts of proofs before autonomous cars become common.
Jobs will go. That's happened with every innovation since the Agricultural Revolution. It isn't imminent though.
Everything we do over networked connections is vulnerable. From Mumsnetting to online banking. This will be no different. Will it be on balance worse than human choices? Who knows?
Worra perhaps the lollipop people should start a campaign for visual braking indicators on the front of autonomous cars too! It sounds like a very good point. Maybe even driven cars

justanother

However in the scenario I suggested where a child ran out into the road and in order to save that child the solution is to drive into oncoming traffic where a bus or lorry is travelling and injuring more people or causing a bigger situation, or go into the parked car possibly injuring yourself the car would make IMO the morally right choice and go into the car 100% of the time, humans probably 50/50.

So my passenger becomes disabled because you have chosen to protect the child? Who pays the cost of that decision? I might prefer to kill the child and would be perfectly able to now with no consequences. Are the insurers/manufacturers who will make this decision happy to pay several millions to my passenger or would they prefer to pay nothing to the child's parents?

carol you neglect the idea of heuristic decision making. The software behind the choices will be far more complex than a flow chart. The decision making process within known parameters will be more focussed than your average human.
The momentum of oncoming vehicles, angles of impact, forces applied to pedestrians, stopping distances. All far more likely to be assessed accurately than a human does.

Giddy I have no doubt it will be more accurate than a human. The point is, what is it trying to minimise? It could choose cost, death, injury, third party involvement - which is the priority and who chooses?