Indeed AI is the capacity of machines to do things that are data based and logic guided devoid of reasoning while human intelligence is a capacity to do things based on data where logic and reasoning are the vehicles but to imagine outside of the system  to explore what is not known or present in our data base of knowledge. For instance the capacity to tolerate fault "fault tolerance" that is included in modern machines is a one step ahead to have machines to reason to make decisions and act accordingly than to follow strict logic rules.  

The primordial question is what is the limit of human intelligence?

It cannot be depressed or overestimate itself :) 

Indeed AI is the capacity of machines to do things that are data based and logic guided devoid of reasoning while human intelligence is a capacity to do things based on data where logic and reasoning are the vehicles but to imagine outside of the system  to explore what is not known or present in our data base of knowledge. For instance the capacity to tolerate fault "fault tolerance" that is included in modern machines is a one step ahead to have machines to reason to make decisions and act accordingly than to follow strict logic rules.  

AI has been the way and shall remain the logical base of any Algorithm that operates human intelligence.

 Hi Yingxu,

You may be interested in Vision-Space. A new from of illusionary space based on perceptual structure (not optical projection). The phenomenon of vision is non-photographically rendered. In Vision-Space (as opposed to picture space) we model visual awareness, the data sets that present and the processes of information exchange that occur over time to assist with the saliency of vision. However, the principle data-set of the phenomenon is a radial field structure that we set out from what we are fixating on. Nothing to do with depth-of-field etc. It provisions 'proximity cues' not depth cues that we have to work out via occlusion cues etc. 

Why is this of interest? Because this field structure present implicit spatial awareness. The space is inherent in the data-set. We attend to this data-set holistically. It's the same structure wherever we look its just what appears there that changes. 

None of our current forms of information display or AI systems work on this basis. We are waiting upon. Our instrumentation is not. This is why Vision-Space is a new form of illusionary space and the technology is categorised as a paradigm shift by the TRIZ process for all forms of information display. 

Implicit spatial awareness plays out in what has been termed 'peripheral vision'. It is in fact contextual vision not a degraded from of central vision! Explicit reality plays out in central vision and is an entirely different animal!

I attach a list of presentation and there are some words on my RG page along with a project.

What is the difference between human intelligence and animal intelligence in relation to AI ?

Hi Malaoui,

Perhaps their perceptual structures that are linked to their biological systems? It's these things that govern our umwelt and that of other animals. I am not a fan of 'intelligence'! That's at best secondary?

What is your definition  of animal?  Is Monkey family included? if so Intelligence should not be used!! The difference therefore shall be cloudy.

Is it really possible for a non-organic entity to "think" at all? Our human intelligence is inextricably predicated on pre-exiting desires, that is expectations rising out of our psycho-physiological processes. Intelligence is organic, "embedded and embodied".

Obviously, there are now "intelligence machines" which manifest  intelligent behavior (AlphaGo recently), yet they are still a far cry from the highly imaginative, autonomous HAL 9000 of the film 2001, A Space Odyssey-- a machine which, in the true sense of the word, can think for itself. 

One thing even a sophisticated AI can't do, I think, is owning up to one's epistemic groundlessness: the fact that at bottom all belief systems are founded upon unjustifiable arbitrary "axioms", foundational premises   whose acceptability depends on anything but their logicality.

A human agent, upon occasionally encountering this groundlessness, would have   ( and usually does not have) no problems dealing with the ensuing epistemic uncertainty; equipped with an animal desire to survive and enjoy certain sense-impressions, the human agent could deal with the logical contradictions in his/her belief-system with recourse to his/her pre-logical desires and feelings. If, for instance, the very act of living on is demonstrated (by some nihilist philosopher) to be logically incorrect! and morally irresponsible, a human being could always retort : " the hell with logic! well, I LOVE life and wanna live to the end and more."

An AI machine can't simply deal with such epistemic breaking points, not without some human-assisted programming it can't. And it simply can't  coherently simulate the human response ( above) since to say of some entity ( the AI here) that it "loves" something is to already acknowledge the presence of a psycho-physiological component ( a cerebellum, for example) within the structure of that entity. 

In other words, how would an AI respond to the question : " ultimately, why do you keep doing the things you do?" Surely, if it responds that it "likes" doing them, it has lied.

Interesting discussion ... There is one aspect I would like to add. Human intelligence is not one thing. According to many researchers, it manifests itself in four different forms: (i) problem solving and behavioral intelligence of human individuals, (ii) specialized collective intelligence of human (professional) teams/groups, (iii) aggregated (social and cultural) intelligence of human communities/populations, and (iv) all-embracing intelligence of man-kind. When we talk about the relationship of natural intelligence and artificial intelligence, not only manifestations (i) and (ii), but also manifestations (iii) and (iv) are to be taken into consideration.

There are many more aspects of human intellect than given in the response of Professor Imre Horvath.  American Psychologist J P Guilford's structure of human intellect comprises up to 180 different intellectual abilities organized along three dimensions- operations, content and products. And, there is no reason to believe this is the final word on human intellect. AI can not handle and predict 'emergence' in large complex systems. 

Obviously we are only at the genesis of this artificial world ... soon we will have researchers who will be able to observe the artificial lives that have been developed as anthropologists and ethnologists observe (have observed) human societies, or ethologists Observe (have observed) the intelligent or non-intelligent behaviors of animals.

AI researchers will move away from the first anthropomorphic tracks to discover other areas of relevance and performance for reasonings and complex problem solving.

Hi Alain,

Surely the parameters of such an artificial word are entirely allocated by the conceptual models in existence at the time of its creation? This would beg the question what would the study of such systems actually mean? This is not a trivial point as without a concurrent study of our relationship with the real and our developing intelligence as complex biological systems, the study of the AI system would be simply entering a hall of mirrors? It would have lost any connection with the all important context. We do this a good deal already and actually in my view, we need less of it not more?

What is a non-anthropomorphic intelligence? Does extraterrestrial intelligence belong to this category? How can (human) AI researchers implement it? Do we indeed need it?

I think it's not very convincing to state what AI cannot do, because we don't even have a good idea of what human intelligence is. The more human intelligence is understood and codified, rigorously, the more AI will be able to achieve.

Most practical implementations of AI today, I mean "practical" rather than for research purposes, are very deliberately limited, to be completely predictable. This is done for the implementation to be safe and therefore usable. Doesn't mean that AI must remain so limited.

Another aspect of this is that sensor technology will continue to improve, which has a direct impact on AI. If human intelligence includes reading "body language," there's no reason to believe that "body language" cannot also be codified, and used by AI. As a matter of fact, AI can be programmed to read facial expressions, and take action accordingly. For instance, camera inside driver-assisted car reads facial expressions of driver, to see if driver is actively engaged in monitoring the control of the car. One can only assume that such algorithms will improve, over time.

Another point is, if the size of recursions, in human intelligence, tends to infinity, one has to wonder what the outcome is? This seems to imply that the outcomes will differ greatly, among individuals. One has to believe that this is where "leaps of faith" come into play, and perhaps even this can be codified.

It's always difficult to make a convincing argument about what cannot be done in the future.

And typically according to predefined or emergent human objectives which may be (very much) influenced by subjective factors ...

AI cannot develop affection.