Dear Bessat:

You are raising an interesting and highly debated question among psychologists, philosophers of mind, neuroscientists, and so forth: "Can an Artificial Intelligence machine, in principle, be 'intelligent' at all?". In addition to this, your 8 points sound good to me.

As you certainly know, there is no universal definition of intelligence. For example, differential psychologists see it in continuous and quantitative terms, that is, how much one is intelligent (see, for this respect, the (in)famous concept of IQ), whereas developmental psychologists see it in discontinuous and qualitative terms, that is, what type of intelligence a given individual has (see, for this respect, Piaget's practical or sensori-motor intelligence; preoperational intelligence; concrete operational intelligence; formal operational intelligence). More to the point, Gardner spoke about 8 forms of intelligence (e.g., spatial, interpersonal. etc.), Baltes spoke about fluid intelligence and crystalized intelligence, and Goldman about emotional intelligence. We can think also of practical and theoretical intelligence

I use to say that there are three types of intelligence: cognitive intelligence, which is related to the true; moral intelligence, which is related to the good, and aesthetical intelligence, which is related to the beautiful. Note that the true, the  good, and the beautiful are universal categories, regardless of how they are seen at different places and times,     

Stephen Hawking's definition of intelligence as " the ability to adapt to change" is also espoused by Piaget. According to Piaget, intelligence is one's capacity for an active adaptation to his/her physical and social words.

All that said, I come back to your question: Can an Artificial Intelligence machine, in principle, be 'intelligent' at all?" As you certainly agree, any answer to this question depends on what conception of intelligence one adopts.

As I see it, intelligence involves, as it were, a grammar or rule-governed component and a semantic or meaning component.  No doubt, a machine, a computer, for instance, has, say,  a grammar, but it is completely blind in terms of meaning and semantics. Because of this, artificial intelligence can be intelligence in cognitive or intellectual terms, but not, for example, in moral, social, or emotional terms.  More to the point, even when machines are intelligent in cognitive or intellectual terms, they obeys to a cognitive program that was developed by humans. So, a machine, however intelligent it may be  is only intelligent in terms of grammar, not in semantic or meaningful terms.

I hope I have  got your question and that this helps.

Best regards

Dear Orlando,

Thanks for the informative response.

What I aspire to do with my tentative definition (above) is capture the "essence" of various manifestations of Intelligence. I believe that almost all the instances of Intelligence involve, in one way or another, perceiving or calculating the relation(s) between (basically) two variables. (Gardner's) Musical Intelligence, for instance, would involve on's cognitive capability in recognizing the similarities, differences or identities between two musical phrases or notes, or even between a musical piece and a thought etc. All intelligence(s) must have something of this sort in common;otherwise, we might be talking about different categories of cognitive aptitudes.

I find your differentiation between the "grammar or rule-governed" and the "semantic" components of intelligence helpful as I think my definition could be also thought of as the deep structure "grammar" of Intelligence in general.

In any case, what I would like to draw attention to is the "temporal" component of the definition: the belief that Intelligence is made sense of only at the background of our time-bound finite life has, I think, devastating implications for the prospect of AI.

My point of departure for my argument, as you have noted, is a universal definition of Intelligence, and the fundamental assumption of my definition is that all forms of Intelligence have furthering of human survival as their ultimate function--I wonder what other possible functions Intelligence might have! 

So, would you say that such a definition, upon further reflection on the multiple instances of Intelligence, serves to capture the essence/grammar of Intelligence? If not, what would be some definitive counterexamples of Intelligence that cannot be accommodated into my definition?More precisely, are there Intelligences that would preserve their intelligenceness in a universe with immortal human beings?

Yes and no. Or - rather not instead of - yes. All machines work on a particular algorithm. It is true that modern machinery can "trained yourself", but even this training is subject to the developed human algorithm. In this sense, if a situation arises that requires a creative solution, the machine will seek a solution envisaged in the algorithm, ie will offer a trivial decision. It is true for example that computers play a good variety of games, even play chess. However, this is not due to their intelligence, and due to this that are fast and can quickly "analyze" and "remember" all possible situations after that select the best decision based on their algorithm.

"Perhaps the greatest significance of the computer lies in its impact on man's view of himself. No longer accepting the geocentric view of the universe, he now begins to learn that mind too, is a phenomenon of nature, explainable in terms of simple mechanisms. Thus the computer aids him to obey for the first time the ancient injunction: "Know Thyself".  -Herbert A. Simon

All algorithms have behind them the knowledge on how we humans do it ourselves. I hope, by the time we learn enough about our own intelligence, we would realize the futility of building such machines.

The problem is that we really do not know what intelligence is. Look at the continuous redefinition of what intelligent behaviour is over time, it was once chess playing and logic. Now it is emotional behaviour, decision making with uncertain knowledge and learning. That is why AI redefines itself over time, a computer beats humans in chess, is the computer intelligent? It seems not. The other problem is that often intelligence is confused with mind and consciousness, often the following equation is assumed intelligence=mind=consciousness, which to me is quite wrong.

@Aleš Kralj

I like Aleš Kralj's new definition of intelligence as: an ability to explore and exploit information. The only thing I would change is: an ability to overcome problems by exploring and exploiting information.

The greatest humans' achievements through intelligence are enabled by humans' skills in logic, synthesis, combination, and imagination. So far, the Artificial Intelligence (AI) turned out to be a formidable competitor to humans in logic and combination skills.

As long as humans preserve their advantages in synthesis and imagination skills, they can utilize AI as a  tool. Once AI level balance in that through exploration and exploitation learning processes, the humans leading position in setting and ovecoming problems might become quite shaky.

From this moment on, machines enabled with AI might want to use humans as a tool. And it is not the worst case scenario.

What I was aiming at with my original question was to embolden the "temporal" component of any form of intelligence. IF  this is demonstrated to be true, then the very idea of intelligence would be unintelligible for an entity ( an AI machine) who simply does not care about time--unless it is programmed to to do so.

"an ability to overcome problems by exploring and exploiting information." L.L. Mizrah

I assume that you (Dear Mizrah) would grant that we aspire to solve "problems" only, in one way or another, to boost our survival, for what else would be the ultimate aim of our problem solving?

The very concept of a "problem" itself is predicated on the paucity of time for mortals like us. As long as whatever we do in the way of solving "problems" (philosophical or otherwise) ultimately serves to enhance and further our survival, we have to grant that "problems" as such exist  and make sense only for such self-consciously survival-thirsty animals as humans. I wonder if ,given infinite time, all our current "problems" wouldn't cease to be perceived as problems at all!

As a neuro-cognitive faculty evolved and perfected over millennia, being "intelligent" presumes both volition and temporarily.  Humans are intelligent because they want to survive and want to survive because they are hardwired to have fears and desires. 

Put otherwise and more logically, the hierarchy of our human logic is such that every "ought" presumes an " if", and every "if" in turn presumes a pre-existing "desire". Even Kantian "transcendental imperatives" presume a willingness to accept the authority of Pure Reason in the first place i.e there is no way to "transcendentally" argue someone into rationality; rather, one first spots a desire in themselves to be rational ( to survive more, in my opinion) and then chooses to kowtow before the "imperatives" of Reason.  

It just seems counter-intuitive how such a system of logic which is both evolved and perfected based on an  animal desire (itself overshadowed by mortality) could be viably transferred unto a non-desiring entity.

Dear Kralj,

Granted, and why does it want  to solve problems? 

Is such an Intelligence transferable to an AI which does not want?

In my opinión, the term is bad employed. In reality, AI considers mathematical methodologies similar as any other engenering área.