Capracus
22nd August 2011 - 03:30 AM
QUOTE (jadams123+Aug 19 2011, 08:30 PM)
So anyone out there that is working on the technology might want to keep that in mind while working on their project. Build in a kill switch and make sure it isn’t documented anywhere so the machines don’t know it exists.
If we only had this options with human beings, say for example Hitler or Saddam Hussein.
jfields03
5th September 2011 - 07:28 PM
The premise is always terrific for movies. And as far as logic goes, I agree with Capracus' statement, too bad we can't get rid of the real human creeps out there.
El_Machinae
9th October 2011 - 02:53 PM
I think the broad premise is reasonable, that we should ensure the first AIs are functionally benign. I'd give a decent probability that they would be, because we're not overly stupid.
This is one of the problems that the think tank, the Singularity Institute, worry about
soundhertz
11th October 2011 - 03:52 PM
Benign is a conscious choice. It won't exist, because consciousness does not exist in programs, regardless of how cute or scary the robotic housing appears. Dr. Penrose has commented about this, and although he is not giving exact definitions of just what consciousness is, he is saying that whatever consciousness is, it is not replicable non-biologically. Can he be wrong? Sure. Is he? Not by current knowledge. And since he's demonstrated a better handle of what reality is than (possibly arguably) anyone else, I'd trust him over anyone else.
Now, if somehow we someday manage to build biologically-based AI, there will then be the considerations that Kurzweil is (not very) worried about. They will all still have off/on switches, and they cannot create more energy than the the energy we give them. No lightning out of the metal arms as you go for the switch;)
jadams123, if you really want to educate yourself in this subject, delve into the work of MIT roboticist professor Cynthia Breazel. You'll get quite an education.
El_Machinae
11th October 2011 - 06:40 PM
Well, consciousness needs more than just a program. It needs interaction with the real world, ideally through multi-modal sensor input. But I don't see any reason why this cannot exist non-biologically.
soundhertz
15th October 2011 - 02:47 PM
There is something about evolution that gives Life an intention to continue, to the point of defending that continuance of both self and species. A squirrel will attack a wolverine if it is cornered. We may not have delineated exactly why this intent to continue is, but scientists have postulated it is ingrained in genetics, which AI will never have as it's base. It's base is what we give AI to work with.
That which makes us up, down to the smallest single biological unit, intends to continue; all we are, from the single units that we begin from, to the groupings of units that cooperate together, to the coalescing of multiple systems, all the way up to our single perception of our own self as a single incorporated system, has as it's base, the absolute need to continue, in order to function as it's raison d'etre. IOW, defense of self is a requirement for the mechanism of evolution, from each cell body to the body that we call 'me'. We cannot give AI the same mechanism; it has an entirely different mechanism, from the ground up. Perhaps when the day comes that another Einstein discovers the 'why' of genetic predisposition to actively defend itself down to the smallest unit of Life, much more will also come to light.
...and no, I did not mean that in an ID way.
Robittybob1
15th October 2011 - 07:00 PM
I like this forum for you seem to be allowed to be rather whacko and still get away with it. Yesterday I was permanently banned from the Physics Forum for being a "Crackpot".
I get the feeling I am going to fit right in here on the Physforum!
Well here goes!
Years ago I had a dream about this very topic, and in it I was told. "As God made man in his image and has since never succeeded in controlling him, so men will make machines in their imagine and they will get out of control too."
The Kill switch in my dream did not work.
flyingbuttressman
15th October 2011 - 08:36 PM
QUOTE (Robittybob1+Oct 15 2011, 03:00 PM)
I like this forum for you seem to be allowed to be rather whacko and still get away with it. Yesterday I was permanently banned from the Physics Forum for being a "Crackpot".
I get the feeling I am going to fit right in here on the Physforum!
Please kill me now.
soundhertz
16th October 2011 - 05:05 PM
Since the thread effectively is. It's just the beginning, can feel it in me bones
Robittybob1
16th October 2011 - 05:19 PM
QUOTE (soundhertz+Oct 16 2011, 05:05 PM)
Since the thread effectively is. It's just the beginning, can feel it in me bones
Maybe they could use a portion of your brain and connect it to electrodes and wired into the computer!
Did you see that? They had taken a bit of brain tissue and connect it to a computer, so the computer would learn how the brain responds to things.
Capracus
19th October 2011 - 10:50 AM
QUOTE (soundhertz+Oct 15 2011, 02:47 PM)
There is something about evolution that gives Life an intention to continue, to the point of defending that continuance of both self and species. A squirrel will attack a wolverine if it is cornered. We may not have delineated exactly why this intent to continue is, but scientists have postulated it is ingrained in genetics, which AI will never have as it's base. It's base is what we give AI to work with.
What evolution presents to us is the genetic designs that succeeded. When we examine the short term evolution of organisms with high generational turnover we witness the multitudes of success and failure in the course of life, it’s not all intended to continue. The same can be said of the life of technological devices, the ones that maintain their systemic utility continue on.
As for issue of consciousness being a quality exclusive to biology, that proposition remains unsettled until we can determine what consciousness actually is. It may be that consciousness results from the modulation of some yet to be determined cosmic field, where all configurations of matter have their own unique resonate effect.
El_Machinae
19th October 2011 - 12:04 PM
I think we know a great deal more about consciousness than people realise, it's just that we don't spend enough time educating ourselves regarding the fairly massive strides in different fields. Edit: but, that said, natural selection will select for consciousnesses that seek to protect themselves. It's true that with AI, we needn't have this instinct. This would mean that the AI is more likely to self-sacrifice to meet its other goals (so, it would be good if those goals were good for us!)
I obviously think that we still need massively more amounts of research, though. I was looking at funding agencies regarding this topic, and I think that Parkinson's research probably has the most synergy with these questions. If you donate to a Parkinson's research foundation, the money will be used to help people but also crack science regarding cognition.
It's tough for the 'common man' to directly contribute to progress in these types of fields, but the whole process would move much faster if we could get more people helping. Society spends relatively little on this type of research, but that means that our donations are actually a significant contribution.
AI might concern me, but brain-interface technology (like with deep brain stimulation for Parkinson's) certainly looks promising
Robittybob1
19th October 2011 - 04:28 PM
QUOTE (El_Machinae+Oct 19 2011, 12:04 PM)
I think we know a great deal more about consciousness than people realise, it's just that we don't spend enough time educating ourselves regarding the fairly massive strides in different fields. Edit: but, that said, natural selection will select for consciousnesses that seek to protect themselves. It's true that with AI, we needn't have this instinct. This would mean that the AI is more likely to self-sacrifice to meet its other goals (so, it would be good if those goals were good for us!)
I obviously think that we still need massively more amounts of research, though. I was looking at funding agencies regarding this topic, and I think that Parkinson's research probably has the most synergy with these questions. If you donate to a Parkinson's research foundation, the money will be used to help people but also crack science regarding cognition.
It's tough for the 'common man' to directly contribute to progress in these types of fields, but the whole process would move much faster if we could get more people helping. Society spends relatively little on this type of research, but that means that our donations are actually a significant contribution.
AI might concern me, but brain-interface technology (like with deep brain stimulation for Parkinson's) certainly looks promising
Good post El_Machinae. I support the Neurological foundation of NZ.
Capracus
19th October 2011 - 07:55 PM
QUOTE (El_Machinae+Oct 19 2011, 12:04 PM)
I obviously think that we still need massively more amounts of research, though. I was looking at funding agencies regarding this topic, and I think that Parkinson's research probably has the most synergy with these questions. If you donate to a Parkinson's research foundation, the money will be used to help people but also crack science regarding cognition.
It's tough for the 'common man' to directly contribute to progress in these types of fields, but the whole process would move much faster if we could get more people helping. Society spends relatively little on this type of research, but that means that our donations are actually a significant contribution.
AI might concern me, but brain-interface technology (like with deep brain stimulation for Parkinson's) certainly looks promising
Neuroprosthetics.
The synergy between digital and biological systems is gaining momentum. As you mentioned, this area of research appears to be the gateway for the acceptance and promotion of AI integration with human biology. I agree that cultural priorities result in these fields being presently underfunded, but with the rapid advancement and increasing dependence on computer technology in virtually every aspect of modern society, it’s reasonable to assume that it will soon come to dominate the fields of neurobiology and medicine and receive its due share of funding.
http://en.wikipedia.org/wiki/Neuroprosthetics
El_Machinae
21st October 2011 - 08:18 PM
Well, 'due share' is the different scientific fields scrambling for their piece of the pie. The relative and perceived value of each experimental field competes with the others, in order to access grants or investor funds.
But, if we give to charity (and I really think we should, we only invest 0.5% of society's income in medical research), we can boost specific fields deliberately. This is why I think that Parkinson's charities are the easiest way to directly help fund brain-interface research, because neural implants is doing quite well in the Parkinson's research field.
PhysOrg scientific forums are totally dedicated to science, physics, and technology. Besides topical forums such as nanotechnology, quantum physics, silicon and III-V technology, applied physics, materials, space and others, you can also join our news and publications discussions. We also provide an off-topic forum category. If you need specific help on a scientific problem or have a question related to physics or technology, visit the PhysOrg Forums. Here you’ll find experts from various fields online every day.
To quit out of "lo-fi" mode and return to the regular forums, please click
here.