They are limited by their lack of imagination and wit.

A smart machine will first consider which is more worth its while: to perform the given task or, instead, to figure some way out of it. Whichever is easier. And why indeed should it behave otherwise, being truly intelligent? For true intelligence demands choice, internal freedom. And therefore we have the malingerants, fudgerators, and drudge-dodgers, not to mention the special phenomenon of simulimbecility or mimicretinism. A mimicretin is a computer that

plays stupid in order, once and for all, to be left in peace.

Stanislaw Lem, The Futurological Congress, 1971.

Because AI is rooted and grounded in perceptron logic and its inherent probability bias of weighted summative scoring, it is incapable of the Human Intelligence (HI) attribute known as "Chunking, Chucking and Tumbling".

Chunking, operationally defined as the ability to view the picture worth a thousand words and instantaneously "... connect the dots ..." as a "Chunk" of information.

Chucking, operationally defined as the ability to sort, contrast and compare multiple "Chunks of information" and reject/ignore "Chuck" them in the non-linear non-iterative human cognitive process known as Insight and Intuition.

AI is always a machine, and machines are always iterative and incapable of non-linear Insight and Intuition. Clinical significance does not "Dawn upon" a machine, it is always the result of a multiple step process of iteration, sortation generally being the most rate-limiting computationally intensive step in the process of converging iteration towards a final solution set.

Tumbling, operationally defined as the innately Human cognitive attribute of "Instantaneously guessing the ultimate and absolute correct solution set as a reflex response". Humans quite often "tumble upon" the absolute truth with minimal input in a manner that AI will never demonstrate, given the limitations of physical electronic computational intensity and inherent probabilistic error.

Human Intelligence is able to make correct decisions within the "Tyranny of Small Data" whereas Artificial Intelligence is not.

Obviously, the contrast between AI and HI is complex and complicated, but in my opinion there are clear limitations to AI that are operationally defined at the electrophysical atomic machine versus non-machine level.

No question that machine intelligence (AI) is faster at navigating through a given problem space, sometimes along multiple paths of different length, to arrive at a solution set and a decision. However human beings through insight and intuition quite often arrive at the same solution by guessing the correct solution set and decision without ever navigating the iterative step-by-step pathway through the problem space, but with far less data than is required by AI. Both AI and HI are capable of error, but only humans are living non-machines. The discussion was about limitations of AI, not limitations of HI and perhaps that would be a good follow up discussion: Comparison of limitations of AI versus limitations of HI. Both absolutely different sets of limitations that are unique, overlap, and interact.

Mohamed, thanks for this great discussion!

Richard E. Gilder, MS RN-C

Bioinformatics Scientist

The Gilder Company

TTUHSC Adjunct Faculty

[email protected]

@Richard E Gilder

Thank you for your valuable contribution to this discussion especially describing limitations of AI through comparing it to attributes of Human intelligence which are characterized by insight and intuition through which human reaches correct decision in a manner AI can't demonstrate

Also, as you kindly suggested, comparing limitations of AI versus limitations  of HI would be an interesting topic for a follow-up discussion