Talk:Artificial consciousness/AI vs AC
What's the difference; should they be merged? --Wikiwikifast 04:31, 23 Apr 2004 (UTC)
- The principal difference is that Artificial Intelligence only aims to simulate human intelligence, while artificial consciousness aims to simulate human consciousness. Therefore the definitions are different, the requirements are different, and in a way the whole context is different. The problem with intelligence is that it is not exactly defined, but bigger problem is that when we try to define it, then the definition often exceeds that what would be necessary only for intelligence. Artificial consciousness is something in what much less people, and only a few AI researchers, are interested in. Artificial consciousness is certainly a separate field, also by far not any (almost no) Artificial Intelligence program is artificial consciousness program, and for example the program what simulates imagination, emotions etc is not Artificial Intelligence program when it doesn't simulate intelligence. There is no consensus of whether artificial consciousness (simulated consciousness, artificial sentience, digital sentience, artificial mind etc) is a field within AI or a separate field. If it is considered to be a subfield of AI, then it must be a separate article, as subfields of AI like Neural Networks are separate articles. When it is a separate field, then it must be a separate article as well. So in either case it must be a separate article and not be merged with Artificial Intelligence. Tkorrovi 15:32, 23 Apr 2004 (UTC)
I understand that AC researchers want to separate themselves from general AI researchers. However, the cognitive science and philosophy of mind academic literature refers to Artificial Intelligence when they probably mean Artificial consciousness. Both the AI and AC articles discuss consciousness, the Turing Test, Searle's Chinese Room argument—it seems quite redundant. In AI, the goal of "strong AI" seems to be the same as that of "artificial consciousness". In addition, philosophical issues of AI would apply to AC, and vice versa. Is there a way to reduce the redundancy? -Wikiwikifast 18:25, 23 Apr 2004 (UTC)
> I understand that AC researchers want to separate themselves from general AI researchers.
- No, not at all, they only don't agree that these two terms are equivalent.
> However, the cognitive science and philosophy of mind academic literature refers to Artificial Intelligence when they probably mean Artificial consciousness.
- This is wrong, and merging AC under AI doesn't help to improve it.
> Both the AI and AC articles discuss consciousness, the Turing Test, Searle's Chinese Room argument?it seems quite redundant.
- Sure it is. I personally think that AI article should not discuss consciousness and AC article should not discuss Turing test, as this was meant to be an AI test and is not proper for AC, especially because it doesn't enable to objectively test certain properties of AC. Searle's Chinese Room argument is often not considered important any more, it is kind of AI versus consciousness, so perhaps must go under AI, or even under consciousness (or mind or philosophy of mind or cognitive science).
> In AI, the goal of "strong AI" seems to be the same as that of "artificial consciousness".
- Yes by some views, more exactly then AC can considered to be strong AI, but is more than artificial intelligence. There is no "theory of strong AI", and maybe cannot be, we may not be able to create strong AI when the aim is only to simulate intelligence. But maybe not. So the question of whether "strong AI" and AC are the same remains open. And if they are the same, then "strong AI" would at least be an inexact term, as it is AC then.
> In addition, philosophical issues of AI would apply to AC, and vice versa.
- Not all issues. There are issues such as prediction, imagination, creativity, emotions etc what apply to AC, but mostly don't apply to AI.
> Is there a way to reduce the redundancy?
- Yes there is. First the AI issues should not be discussed under AC. There are many of them at present, with what I personally mostly don't agree, or don't agree that they are appropriate for AC. If you like, you may help to edit AC article to remove these issues or refer to AI article instead of discussing them there. I don't much like to do it at my own discretion only. And second, maybe also such issues like consciousness should not be discussed under AI, there are plenty of places like consciousness, mind, philosophy of mind and cognitive science, to discuss these. Finally, AI is defined through intelligence in the beginning of AI article, and it is everywhere defined so, also because of that alone, it would not be possible to put AC into the context of AI article. AI can be put into context of AC as a special case, but I suppose that it was not meant to merge AI article under AC.
- Tkorrovi 19:54, 23 Apr 2004 (UTC)
"Conscious but stupid" in a phrase explains why AC and AI are not equivalent. Paul Beardsell 02:00, 26 Apr 2004 (UTC)
- Very good Paul, we don't agree in so many things, but we agree in this, it indeed determines extremely well what is the difference. Tkorrovi 19:47, 26 Apr 2004 (UTC)
No offence, but who besides you, Tkorrovi, and this one "Professor Igor Aleksander" uses the term 'artificial consciousness'? I agree the term is much clearer, but even Daniel Dennett and others would say they study AI, not AC. --Wikiwikifast 04:30, 26 Apr 2004 (UTC)
- The AC article also covers simulated consciousness, machine consciousness, artificial sentience, digital sentience (what was a separate article before, but recently merged under AC by request of Raul654) etc. Some consider them as synonyms, others consider that AC encompasses all these. There are many papers (such as university publications) written in these fields by professional AC researchers, almost no peer-reviewed papers though, but it is said in Wikipedia help pages that "you don't have to get all of your information on entries from peer-reviewed journals". I am independent freeware programmer and study it only because of interest, by education I am Automatic Control engineer. Why Daniel Dennett says that he studies AI, I don't know, he is an expert what concerns consciousness, others are taken other approach and say that they study AC, machine consciousness, digital sentience etc. AC article was first created under Psychology, then emptied because of lack of content, and then started again by me. Digital Sentience article was created somewhere in Computer Science hierarchy. There is obviously some interest concerning the topic, but it is of course not as big as concerning AI. It is a field of study, and as such I think must be in Wikipedia, this makes Wikipedia also a bit richer than other encyclopedias, where you cannot find what is AC or its synonyms. Tkorrovi 18:52, 26 Apr 2004 (UTC)
- I think it's not exactly obvious that Daniel Dennett uses the term AI in a sense of AC. He recognizes the Turing view that AI might not be the way the brain does it in many regards. But he considers AI as something necessary to understand certain functions of the brain, like language processing, or, holding a top down approach, as a kind of first step towards AC. Why he don't study AC may be because the exact research he does in AI concerns modelling the mental abilitities what go under intelligence Tkorrovi 09:05, 27 Apr 2004 (UTC)
- Daniel Dennett wrote (the possibly arrogantly titled) Conciousness Explained so let's be clear: Consciousness is a central concern of Dennett's. In that book one of the many things he does is to consider how consciousness might be artificially implemented so as to illustrate human consciousness. So AC is a concern also of Dennett's. Dennett, like many Strong AI proponents, argues there is no obstacle to AC arising from AI. So much for that. Also I am unfamiliar with the Turing view as expressed in the above paragraph. Paul Beardsell 12:59, 27 Apr 2004 (UTC)
There is Tkorrovi's AC forum the URL of which I am sure he will give you but as far as I can tell none of the contributors are AC researchers. Doubtless this will be seen as a mean personal attack. But there is lots of research into AC and the best quick resource I can offer you is this
Google search. Recent exhortations to my fellow editors of artificial consciousness on that talk page are to concentrate on representing the established positions of respected research and to stop making it up. Paul Beardsell 04:55, 26 Apr 2004 (UTC)
The most applicable Turing result to strong AI and to AC is that all computing machines are equivalent in capability except for speed and memory capacity (Church-Turing thesis). The question arises: Is the brain-mind merely a machine? If so, and if Consciousness is a capability of the brain hardware-software machine, then genuine/true/real consciousness implemented in an artifact (i.e. genuine/true/real AC) is a possibility. Some would say it's inevitable. Thereby I hope to have shown that AC is worthy of serious consideration and its own article. Paul Beardsell 12:59, 27 Apr 2004 (UTC)