Advanced Driving Hub

Summary here:
http://www.technologyreview.com/view/54 ... d-to-kill/

Original paper here:
http://arxiv.org/pdf/1510.03346v1.pdf

the ethics of an autonomous car - should it kill you the driver to save 10 pedestrians? if so, would you buy it?

Alasdair

Why do we think self-driving cars are inevitably going to crash? Surely the whole point is not to get themselves into these situations in the first place?

maybe we don't, but it would be a very confident person who believed nothing could go wrong - we have already seen with the Google cars that they take such a cautious stance regarding possible issues (including waiting for pedestrians on the pavement to move away, even though of no consequence) that other drivers drive into them! On the basis that AI as a concept is still a long way from reality, can we believe that a computer will have the ability to make the same / right / better judgement than a human? It can't on an AI basis, so it has to have every eventuality programmed into it - which means you have to program scenarios such as these - which leads to an ethical dilemma...

Alasdair

On a related note, what about if the hackers change the ethics....
http://www.eetimes.com/document.asp?doc_id=1328103

Computer gamers of a certain age will be familiar with what could happen...

Every now and then there is a fad. We had 3D TV for instance, and before that true AI. But now we are starting to see that technological progress follows its own broken master plan.
The same, I believe, applies to autonomous vehicle. The story is a tad long, but bear with me.

I worked about 9 years ago on the ethics of nanotechnology. Back then, nanotechnology was all the rage. Well,...we know what happened since then (or rather we don't): nanotechnology didn't deliver the all-encompassing changes that some anticipated. Anyhow, at the time, one interesting question was about the relationship between consciousness and technology: what would happen if you replaced brain cells, one by one, with their computerised equivalent, performing the same function? Wouldn't you obtain a thinking machine? At the heart of it, the debate was about whether we, humans, were anything more than biological machines, soon to be made obsolete by nano-manufactured copies of our ourselves. Of course, we now know that there are fundamental reasons why this won't happen any time soon (but this is an entirely different and fascinating topic).
However and without going into too much details, I believe that this mechanistic view of the world results from top-down ignorance, that is, ignorance of the finer details of life and reality.
The concept of an autonomous car, just as the concept of humans as biological machine, is easily grasped and feels intuitive for many and even self-evident to a few. However, look more closely and you are assaulted by messy details: thorny ethical issues, unforeseen events, bugs, hacks and politics, cancers and self-contradictions. This messiness, in my view, IS the reality, as opposed to our simplified top-down generalisation. Unfortunately, scientists and engineers and especially young scientists and engineers, tend to lack the experience of life required to understand that the devil, as they say, is often in the details.

Coding a car that is proficient in 99% of situations was always going to be the easy bit.

Wake-up, dreamers, and taste the red pill.

ChristianAB wrote:Every now and then there is a fad. We had 3D TV for instance, and before that true AI. But now we are starting to see that technological progress follows its own broken master plan.
The same, I believe, applies to autonomous vehicle. The story is a tad long, but bear with me.

I worked about 9 years ago on the ethics of nanotechnology. Back then, nanotechnology was all the rage. Well,...we know what happened since then (or rather we don't): nanotechnology didn't deliver the all-encompassing changes that some anticipated. Anyhow, at the time, one interesting question was about the relationship between consciousness and technology: what would happen if you replaced brain cells, one by one, with their computerised equivalent, performing the same function? Wouldn't you obtain a thinking machine? At the heart of it, the debate was about whether we, humans, were anything more than biological machines, soon to be made obsolete by nano-manufactured copies of our ourselves. Of course, we now know that there are fundamental reasons why this won't happen any time soon (but this is an entirely different and fascinating topic).
However and without going into too much details, I believe that this mechanistic view of the world results from top-down ignorance, that is, ignorance of the finer details of life and reality.
The concept of an autonomous car, just as the concept of humans as biological machine, is easily grasped and feels intuitive for many and even self-evident to a few. However, look more closely and you are assaulted by messy details: thorny ethical issues, unforeseen events, bugs, hacks and politics, cancers and self-contradictions. This messiness, in my view, IS the reality, as opposed to our simplified top-down generalisation. Unfortunately, scientists and engineers and especially young scientists and engineers, tend to lack the experience of life required to understand that the devil, as they say, is often in the details.

Coding a car that is proficient in 99% of situations was always going to be the easy bit.

Wake-up, dreamers, and taste the red pill.

Here here.

Nigel.

fungus wrote:Here here.

Where? Where?

So, CAB, are you saying that programming an auto car might be a bit 'hit and miss'?

Strangely Brown wrote:

fungus wrote:Here here.

Where? Where?

There there.