Interesting article by Bernardo in the Scientific American (or at least the SA web site):
https://blogs.scientificamerican.com/obs...l-follies/
Also interesting that he uses the kidney analogy - I wonder whether he's been browsing this thread as some of his arguments seem very pertinent to this discussion.
Quote:After all, the mere emulation of a phenomenon isn’t the phenomenon: I can emulate the physiology of kidney function in all its excruciating molecular details in my desktop computer, but this won’t make the computer urinate on my desk.
I had to smile about his observation about the current vogue for AI in SciFi - it is something I have been increasingly aware of lately and, of course he is right: Artificial Intelligence does not equate to artificial consciousness - which is the assumption made in many of these stories. I'm watching the new Netflix show, Altered Carbon (I also read the book years ago) where the idea is taken to the extreme.
I had to smile about his observation about the current vogue for AI in SciFi - it is something I have been increasingly aware of lately and, of course he is right: Artificial Intelligence does not equate to artificial consciousness - which is the assumption made in many of these stories. I'm watching the new Netflix show, Altered Carbon (I also read the book years ago) where the idea is taken to the extreme.
Though Karmy wont see this reply others will.
It is wise to think the bold is provisional.
(2018-03-07, 08:19 PM)Steve001 Wrote: [ -> ]I had to smile about his observation about the current vogue for AI in SciFi - it is something I have been increasingly aware of lately and, of course he is right: Artificial Intelligence does not equate to artificial consciousness - which is the assumption made in many of these stories. I'm watching the new Netflix show, Altered Carbon (I also read the book years ago) where the idea is taken to the extreme.
Though Karmy wont see this reply others will.
It is wise to think the bold is provisional.
Why? I guess if you are redefining "wise" to mean "deriving from faith" then I could see how one might make such a statement. Otherwise, the "wise" position would be to make no such assumption considering where our current scientific understanding of consciousness sits.
(2018-03-07, 09:00 PM)Silence Wrote: [ -> ]Why? I guess if you are redefining "wise" to mean "deriving from faith" then I could see how one might make such a statement. Otherwise, the "wise" position would be to make no such assumption considering where our current scientific understanding of consciousness sits.
I just find it utterly remarkable how a simple word, in this case "wise" is misconstrued. No wonder we sometimes talk pass each other. It means:
having or showing experience, knowledge, and good judgment.
My comment was meant as sarcasm. You did notice the word "provisional I'm sure. His statement had a tone of it ain't never gonna happen. But here's the rub. Nobody, Including Karmarling knows how far AI will advance. Perhaps to the point where artificial AI does equal artificial consciousness.
I think the point about AI as it is currently understood and implemented, is that it isn't mysterious, it is just using logic circuits to implement a set of rules. Its basis is the same as an alarm clock, using hardware to implement a rule. There's no consciousness lurking in that alarm clock.
(2018-03-08, 04:11 AM)Typoz Wrote: [ -> ]I think the point about AI as it is currently understood and implemented, is that it isn't mysterious, it is just using logic circuits to implement a set of rules. Its basis is the same as an alarm clock, using hardware to implement a rule. There's no consciousness lurking in that alarm clock.
We'll all have to remember this next time someone cheers for panpsychism.
(2018-03-08, 02:10 AM)Steve001 Wrote: [ -> ] I just find it utterly remarkable how a simple word, in this case "wise" is misconstrued. No wonder we sometimes talk pass each other. It means: having or showing experience, knowledge, and good judgment.
My comment was meant as sarcasm. You did notice the word "provisional I'm sure. His statement had a tone of it ain't never gonna happen. But here's the rub. Nobody, Including Karmarling knows how far AI will advance. Perhaps to the point where artificial AI does equal artificial consciousness.
Invoking sarcasm now? That's a new one. Remarkable indeed. My apologies.
A couple of follow up points. I think Bernardo was pointing out, as Typoz has above, that intelligence is not the same as consciousness. The argument he was making was that we don't even understand what consciousness is, let alone understand what gives rise to it. We do, however, understand how to emulate intelligence - at least some kinds of intelligence. There's the simple mechanistic example of the clock given by Typoz and also the complex algorithms and logic circuits involved in AI.
This latter example mimics, and often exceeds, human intelligence when it involves calculations. But the process does not involve consciousness other than the consciousness of the human designers and programmers. The process is mere switch flipping which is all that is happening in logic circuits. This, in no way, equates to the human conscious experience which involves subjective appreciation. Calculation by flipping switches does not result in subjective experience. That is where the category error is made by those who assume that consciousness can be reduced to something else - either a mechanistic cause with neurones substituted for logic gates or the panpsychist assumptions described by Bernardo where they reduce consciousness itself to "pixels of experience".
(2018-03-08, 06:11 PM)Kamarling Wrote: [ -> ]A couple of follow up points. I think Bernardo was pointing out, as Typoz has above, that intelligence is not the same as consciousness. The argument he was making was that we don't even understand what consciousness is, let alone understand what gives rise to it. We do, however, understand how to emulate intelligence - at least some kinds of intelligence. There's the simple mechanistic example of the clock given by Typoz and also the complex algorithms and logic circuits involved in AI.
This latter example mimics, and often exceeds, human intelligence when it involves calculations. But the process does not involve consciousness other than the consciousness of the human designers and programmers. The process is mere switch flipping which is all that is happening in logic circuits. This, in no way, equates to the human conscious experience which involves subjective appreciation. Calculation by flipping switches does not result in subjective experience. That is where the category error is made by those who assume that consciousness can be reduced to something else - either a mechanistic cause with neurones substituted for logic gates or the panpsychist assumptions described by Bernardo where they reduce consciousness itself to "pixels of experience".
Brain function appears to be more analog than digital so I agree, thinking about switch flipping is unlikely to get us very far.
Re: The kidney.
IMO, Bernardo dodges the point by making the kidney model a computer one, and then he appears faux indignant that the kidney won't urinate on his desk. This is a weak argument.
A physical model of the kidney could be made to function and pee on his desk (
https://www.meddeviceonline.com/doc/worl...ls-by-0001 for example). What the kidney needs is inputs (water and the stuff to be filtered/excreted).
A human brain also (appears to) need inputs to develop sensory perception and ultimately, a conscious awareness of its environment. As I've said before, the problem with AI is that it has, in some ways, to be made a bit crappier to give us the fallibilities that make us human. The experiences that shape us: the good choices, the bad choices, success and failures... Competition, rejection, connections, love and loss etc etc etc etc ..... The neuroses, pettiness, hang-ups and uncertainties that result from all of this.
Any accurate human AI also has to have a way more fluid, unreliable and malleable memory system.
I'm actually with Bernardo, I'm doubtful 'real' AI it can be achieved (but for different reasons).