Current Affairs Robotics and AI....

Status
Not open for further replies.
peteblue said:
But it will happen Bruce, it may take another 50 years but it will happen......
Click to expand...

Hard to say isn't it? I don't doubt that things like machine learning will have an impact upon society. I'm very much in the camp that believes we can adapt and that the technology will be hugely beneficial to the world.

Will we ever achieve an artificial general intelligence? I'd say that's anything but certain and might even be up there with the likes of fusion power. Don't you think it's odd that in the 18 months since DeepMind began working with Moorfield's on retinal imaging that there hasn't been a commercial rollout of the technology, either in Moorfields itself or in any other eye hospital or unit in the world?

I think that, if nothing else, should cause us to view with caution predictions that the 'robots will take our jobs' within the next few years.
 
Bruce Wayne said:
Hard to say isn't it? I don't doubt that things like machine learning will have an impact upon society. I'm very much in the camp that believes we can adapt and that the technology will be hugely beneficial to the world.

Will we ever achieve an artificial general intelligence? I'd say that's anything but certain and might even be up there with the likes of fusion power. Don't you think it's odd that in the 18 months since DeepMind began working with Moorfield's on retinal imaging that there hasn't been a commercial rollout of the technology, either in Moorfields itself or in any other eye hospital or unit in the world?

I think that, if nothing else, should cause us to view with caution predictions that the 'robots will take our jobs' within the next few years.
Click to expand...

Oh, I don’t go along with a short term techno/social change, but I do believe that the enabling technology and the AI will develop massively. History shows that technology and its uses are on a steep upward curve. So, on the assumption that it will happen, we need to be asking the right questions now about how we embrace it and utilise it for humanity. If we don’t then industry will do it for us and not necessarily for everyone’s benefit. ‘Can we adapt ?’ is one of those questions, and I think the answer may well be ‘not all of us’, that’s when it will get tricky........
 
peteblue said:
Oh, I don’t go along with a short term techno/social change, but I do believe that the enabling technology and the AI will develop massively. History shows that technology and its uses are on a steep upward curve. So, on the assumption that it will happen, we need to be asking the right questions now about how we embrace it and utilise it for humanity. If we don’t then industry will do it for us and not necessarily for everyone’s benefit. ‘Can we adapt ?’ is one of those questions, and I think the answer may well be ‘not all of us’, that’s when it will get tricky........
Click to expand...

Well that is the concern isn't it? Automation is quite likely to occur in lower skilled tasks first, but there's a lot to suggest that the lower skilled folk are the worst at adapting (or learning new skills if you like). The Government Office for Science did some work on this a month or so ago (https://www.gov.uk/government/uploa...esight-future-of-skills-lifelong-learning.pdf if you want to read it), and it doesn't make great reading.

It found that those who do engage in education as adults tend to be wealthier and come from a high existing skill level. Those with fewer qualifications to begin with would often cite barriers such as a lack of confidence, lack of interest and a sense that they're too old.

So it's not a lack of money or lack of opportunity that is really the issue, and as such it's going to be a much tougher nut to crack than just throwing money at it.
 
Bruce Wayne said:
Well that is the concern isn't it? Automation is quite likely to occur in lower skilled tasks first, but there's a lot to suggest that the lower skilled folk are the worst at adapting (or learning new skills if you like). The Government Office for Science did some work on this a month or so ago (https://www.gov.uk/government/uploa...esight-future-of-skills-lifelong-learning.pdf if you want to read it), and it doesn't make great reading.

It found that those who do engage in education as adults tend to be wealthier and come from a high existing skill level. Those with fewer qualifications to begin with would often cite barriers such as a lack of confidence, lack of interest and a sense that they're too old.

So it's not a lack of money or lack of opportunity that is really the issue, and as such it's going to be a much tougher nut to crack than just throwing money at it.
Click to expand...

I’ve skimmed through it and will read in detail tomorrow, however this seems a pretty standard learning and development report. It gives a nod towards technology but really only towards automation, without really analysing the potential effect of AI. This is reflected in the charts showing the effects upon manual labour without real consideration about the effects upon professional groups. There will come a time where either humans adapt the technology to enhance their own capabilities, under our control, or humanoids develop and control mankind utilising that spark of thought that cannot be programmed. It may well be that those who are currently regarded as being below average intelligence master this approach, as a child masters today’s technology, whilst the current professionals and intelligentsia find themselves left behind......
 
Tim Urban's waitbutwhy piece on AI is a very good read - hugely speculative but that is unavoidable with this subject. Even if you think he's wildly off-beam he's very good at putting human progress in context. As he says - the magnitude of progress between earth circa 2015 and 2030 could and is even likely to equate to that from the 1750s to 2015, such is the exponential nature of human endeavour. Mind-blowing if true, as imagine what someone from the early days of the industrial revolution would make of today. That will be us in 15 years time.
 
Black Belt Jones said:
Tim Urban's waitbutwhy piece on AI is a very good read - hugely speculative but that is unavoidable with this subject. Even if you think he's wildly off-beam he's very good at putting human progress in context. As he says - the magnitude of progress between earth circa 2015 and 2030 could and is even likely to equate to that from the 1750s to 2015, such is the exponential nature of human endeavour. Mind-blowing if true, as imagine what someone from the early days of the industrial revolution would make of today. That will be us in 15 years time.
Click to expand...

Excellent article. Our problem is that we humans don’t progress exponentially, our thinking is sometimes limited by what we know or what we are told. To detach ourselves from today’s reality and imagine what may be is the real test. Of course you then have to be able to understand how to operate or even use or appreciate what we have. We all expect our smartphones to talk to us, but I’m now having conversations with my car and it doesn’t seem right......
 
rascal said:
AI early diagnosis could save heart and cancer patients

http://www.bbc.co.uk/news/health-42357257
Click to expand...

It's interesting stuff, but it's some way short of what Pete was saying in his original post. It's essentially training an algorithm to spot early signs of particular diseases on scans. There have been numerous projects of this ilk for other diseases and isn't a huge leap forward in terms of the underlying technology. I also know from personal experience that getting this technology (or any technology tbh) into the NHS is a painful and disjointed process. It will be years before even something as relatively straight forward as this is available in every cardiology department in the country.
 
Bruce Wayne said:
I think it's indicative that despite DeepMind gaining a lot of publicity for their Go project, they haven't actually got a single commercial application out of their work (they do a bit with Google data centres and have had a couple of pilot projects that haven't been followed up on). It's largely the same with IBM, who despite Watson gaining a huge amount of publicity after winning Jeopardy has yet to add much to the bottom line of the company, much less end the record run of declining quarterly revenues.

What there are is a huge number of very interesting pilot and research projects out there, but as I'm sure given your experience in business you know, the road from pilot to fully fledged, mass market product is treacherous and far from certain. We're getting massively carried away, largely as a result of people with little real knowledge or experience in the sector or with the technology losing their heads. The fundamentals of AI aren't developing at a rate of knots at all. What is largely behind the progress that has been made is a huge amount of available data, and more powerful computers to process that data. We haven't had tremendous breakthroughs in the underlying AI, despite the hoopla suggesting otherwise, and we're still an awful long way from any kind of general intelligence.
Click to expand...


That may be true at the moment, but it's the kind of thing that could sneak up on us very quickly. Technology is developing at exponentially faster rates, each step forward quicker than the last. I'm not saying this is going to lead to a robot dictatorship, just that it is well worth thinking about the potential ramifications of it now and preparing for it, as humans have had a historically patchy record adapting to new and extreme changes. Interpreting current and future rates of development with old models would be a mistake imo
 
Prevenger17 said:
That may be true at the moment, but it's the kind of thing that could sneak up on us very quickly. Technology is developing at exponentially faster rates, each step forward quicker than the last. I'm not saying this is going to lead to a robot dictatorship, just that it is well worth thinking about the potential ramifications of it now and preparing for it, as humans have had a historically patchy record adapting to new and extreme changes. Interpreting current and future rates of development with old models would be a mistake imo
Click to expand...

Well there's a whole bunch of stuff the NHS should be doing to better cope with the technological changes occurring, but it's something they've never been very good at, and are only really getting worse as the pace of change quickens. There's an awful lot of talk and not a lot of action.

As for exponential rates of change though, that is largely founded on Moore's Law. If we apply that to the example in the BBC article, it might make the machine quicker at diagnosing things, it might make it more accurate at doing so, but it won't make the machine conscious that it's doing it or able to do it of its own volition. That requires a lot more than just raw computing power.

There's a danger that we get so bogged down with very philosophical issues and don't give attention to quite mundane, yet hugely important issues. For instance, data governance isn't very sexy, but it's fundamental to AI, yet the NHS blunder consistently on this topic and certainly haven't shown any leadership since care.data was squashed a few years ago.
 
Prevenger17 said:
That may be true at the moment, but it's the kind of thing that could sneak up on us very quickly. Technology is developing at exponentially faster rates, each step forward quicker than the last. I'm not saying this is going to lead to a robot dictatorship, just that it is well worth thinking about the potential ramifications of it now and preparing for it, as humans have had a historically patchy record adapting to new and extreme changes. Interpreting current and future rates of development with old models would be a mistake imo
Click to expand...

A few pieces here from folks that know more about this than you and I :)

https://www.technologyreview.com/the-download/609875/the-case-against-deep-learning-hype/
 
Status
Not open for further replies.

Welcome

Join the Everton conversation today.
Fewer ads, full access, completely free.

→ Log in 🔑 Register
🛒 Visit Shop

Support Grand Old Team by checking out our latest Everton gear!
Back
Top