This article is from the source 'guardian' and was first published or seen on . It last changed over 40 days ago and won't be checked again for changes.

You can find the current article at its original source at http://www.guardian.co.uk/commentisfree/2013/apr/02/is-the-future-robotic

The article has changed 6 times. There is an RSS feed of changes available.

Version 2 Version 3
Is the future robotic? Find out at midday Is the future robotic? Find out at midday
(about 3 hours later)
In the Guardian Oxford London lecture last month, professor of robotics Paul Newman sought to dispel some of the myths about his field. Were you cheated by visions of K9 and C3PO? Watch the video above for some insight into what robotics really means, and what the implications are for our technological future.In the Guardian Oxford London lecture last month, professor of robotics Paul Newman sought to dispel some of the myths about his field. Were you cheated by visions of K9 and C3PO? Watch the video above for some insight into what robotics really means, and what the implications are for our technological future.
And on Wednesday at midday Newman will be online here to answer your questions about machine vision, robot cars and whether society is ready for technology that can save lives – or endanger them.And on Wednesday at midday Newman will be online here to answer your questions about machine vision, robot cars and whether society is ready for technology that can save lives – or endanger them.
Post your questions below from 10am (UK time) on Wednesday.Post your questions below from 10am (UK time) on Wednesday.
CaptCrash
Serious question.
Robotics can do for consistantly predicatble physical processes, what computers can do for consistantly predicatble mental processes, although I do accept that some aspects of AI may mimick the problem solving that inconsistancy brings up.
My question is, when robots and computers can do for the wealthiest in humanity what the wealthiest formerly employed the poorer to do, what space is there for the poors existence?
Paul Newman replies:
I think technology, without doubt has improved humanity's standard of living and will continue to do so. Just look at health care - computing and robotics I hope will be no different. That's my take any way
Iamyourhuckleberry
Can you give us a precise definition of what constitutes a more or less intelligent cybernetic mind - system?
I mean, I am not asking for a reply from Alan Turing, but perhaps you could have a go.
Paul Newman replies:
This is really interesting - my take is that of an engineer. If we can write down what success looks like, crisply, we can expect every chance of making progress. To me the problem with general AI comes from not actually knowing what we want to build.
clonelordshin
4 questions:
Which method do you think is most likely to create machine intelligence greater than our own? Will it be evolved or designed? (and no, the're not the same thing) Do you think it will be potentially harmful to humans, computers, etc? And what would be the hardest human ability to program into a computer?
Paul Newman replies:
I think design is a form of evolution. Every day computers are getting more useful and we are learning more about how to use them. We are layering complexity bit by bit. I see smart machines evolving by sequentially solving clearly stated problems - then one day we will turn around and say "how remarkable -that machine is so complex surely it could not have been designed with such complexity...whats going on?"
Regarding your second question - most things can be harmful if used to harm. I do not think robotics or machine intelligence is intrinsically harmful and I do worry that this is something worry about. I'd like to address that.
TheMackenator
Aside from machine vision and similar problems that are basically technical, what challenges do you see in getting machines to think like / behave like humans? What crucial differences are there in the "thought processes" of each?
Paul Newman replies:
I think its *all* a technical problem. If that were not so then there must be something spooky or un-makable about intelligence and that seems really hard to believe. I'm not trying to cop out here - really we are created from a code and a stream of experiences - yes its extremely complicated but it is just information, physics and chemistry. Did you question hint that you are wondering if there is something non-technical about intelligence? I'm interested in your view...
SBY818
How much consideration is being given to the accountability of autonomous robots? Taking the driverless car example, even if road accident rates go down as a result of widespread use, the first accident will undoubtably be blamed on the robot itself, and hence the manufacturer, which will vastly limit progress.
I guess what I'm saying is, are there currently any sort of systems or protocols being considered as to define ownership and liability for autonomous robots? I think this could make a big difference to the rate of progress.
Here are Paul Newman's answers to your questions:
Paul Newman replies:
There is lots of work on accountability and its a really important area. I can offer my own thoughts - we have well established engineering techniques to do the best we can to make sure things are safe. We can't prove that the engine won't fall off the wing but we can do a whole load of work to ensure that the risks are understood and managed and furthermore that we have the tools to do a forensic examination of what actually happened if/when something terrible happens like a plane accident. I think we should look to the same professionalism when understanding how to give people the autonomous systems they crave.
Sealion
Bearing in mind that any advances in self driving cars are going to be co-opted by the military and put into self flying drones, do you think we should be looking at the legal and ethical aspects of robotics development now, rather than later?
Related to that, what are the robotics community's take on Asimov's 3 Laws?
Paul Newman replies:
Back in 2010 I was involved in an event on this - you can read more here.
And there are many very wise people thinking about this - sometimes in a jaundiced way sometimes in a balanced way. My ongoing worry though is we throw away the baby with the bath water and focus too much on what is bad and not what amazing good could happen. If silicon chips were being invented now would we worry they could be co-opted for ends we disagree with? Yes we might, but we would certainly also be talking about all the good stuff. I know I'm boring about this point - but I do want to hear the good and the bad.
But to hit your main point again - Its extremely important to think about the ethics just like it is for all new emerging technologies.
wanderingalbatross
Do you think the obsession with machines becoming conscious, taking over the world and destroying humanity (in hollywood, the media, sci-fi) detracts from real discussion about the future of robotics?
Paul Newman replies:
Yes - massively.
I would be delighted and stunned to have a machine with the cognitive ability of a commonplace animal - such as hedgehog. The point is we are *so* far away from any sort of science-fiction-level intelligence it (some sort o technological armageddon) is not something we should be overly worried about right now.
There are bigger issues like safety, trust, cost, economic impact and of course the engineering know-how which we must keep a firm eye on.
jbshort4jb
In light of recent funding announcements re: Obama $100m brain project and European Markram Human Brain Project (HBP) €1.1b over 10years, what are the prospects for UK funding in the robotic sphere?
Paul Newman replies:
Good! look at this. The robotics share of this at least (thats all I am qualified to comment about) will create jobs and I hope a sovereign capability in autonomy - this is a technology that is coming and we should be wanting to lead as a nation.
Sealion
Outside of a controlled environment, how certain can a robot be? If, as you say, we cannot program for every eventuality and must teach robots to learn, analyse and make decisions, how hard is it to reach an acceptable threshold for safety in something like an automatic car? Do you have any comparable data for how certain a human is in similar situations?
Paul Newman replies:
Fabulous question. The way we deal with this is to explicitly deal with uncertainly at every step - from sensor noise, to training data - everything is processed as being "representative yet not certain". This allows decisions to be made as a function of certainty (or uncertainty).
Now what constitutes "acceptable" is something society can decide upon - in exactly the same way as we decides what an acceptable failure rate on tyres. At the end of the day we get to choose what is safe - if we don;t like the risks, we won't buy it or we will legislate against it. Society decides what tech it wants because of the market. I'm positive like that.