Thursday, December 16, 2010

'Artifice'ial Intelligence

Artificial intelligence is an oxymoron (I love oxymorons - they are awfully nice!). What is artificial cannot be intelligent, and what is intelligent cannot be artificial (in my humble opinion). I am sure this will invite violent disagreement from different quarters, dimes, pennies and cents (though 'it all makes perfect sense'), but I don't think the common man is terribly excited about change (pun intended) towards a future with synthetic beings posing as sentient beings. Movies like The Matrix perpetuate that discomfort while we hurtle towards the technology-driven futurist-predicted singularity event in 2045 when apparently computer-based intelligences will significantly exceed the sum total of human brainpower (more on that in a separate article in the works - Cyberia: The Qwest for Cingularity).

Intelligence is an extension of the consciousness, and the homo sapiens species has been blessed with a level of intelligence much higher than other life-forms (the quotient varies, but the base level itself is significantly high). It allows us to deduce, assess, analyze, extrapolate, and sometimes see around corners. Robots would need a horizontal periscope to be able to see around corners. The Citibank CEO Vikram Pandit was famed for being able to see around corners. Too bad he did not see himself being cornered even after seeing around 3 corners to his right (this is like 3 rights making a left) and seeing himself seeing around corners! Reminds me of the scene in the movie 'Being John Malkovich' where John Malkovich gets inside his own head and sees all people as Malkoviches, speaking in sentences comprising repetition of a single word - 'Malkovich'.

Intelligence allows us to take parts and build something significant bigger than the sum of its parts, and the key thing is that it cannot be measured in teraflops. It is that intangible, unquantifiable capability not too different from the befuddling concept of time before the big bang. What needs to be recognized is that a man-made construct will be based on some fundamental rules - you can make the rules as complex as possible and design a self-learning neural network, but at the heart of it are rules that cannot be violated. The artifice of that logic or entity would be that it could fool a human being into thinking that it made a decision which was completely outside of those rules, but deep analysis would reveal that it was made per a decision tree so complex that the navigation of that branch could not be predicted upfront. The Deep Blue computer made an entirely unexpected move which resulted in Kasparov's defeat - it was not dissimilar to something a human being would make, but the fact that this move could not be predicted by human beings did not automatically invite claims of sentience in that computer. Isaac Asimov predicted a robot-infested future where robot psychologists would be needed to work with robots having frozen positronic pathways resulting from fatal indecision (this is interesting - I am using a word that would typically be associated with human beings - that word is fatal - am I a closet artificial intelligence supporter?) associated with conflict across his brilliant 3 Laws of Robotics.

Intelligence is organic to a life-form. It cannot be inorganic. Drawing chemistry parallels, we are talking about inorganic entities (silicon-based) constructed by an organic entity trying to match and even exceed the intelligence of that organic (carbon-based) entity. We need to keep in mind that intelligence is not about processing power - it's about making those irrational yet successful quantum leaps that even kangaroos cannot.

Synthesis of silicon into sentience is scientific subterfuge and self-deceit, simply stated (an alliteration - need to explore this further).

So artificial intelligence is an oxymoron. It's actually intelligent artifice, an interesting phrase considering the embodied redundancy - intelligence is implicit in artifice. Maybe the people who coined the term 'artificial intelligence' recognize that it is artificial (note that they do not say synthesized intelligence, which would convey something entirely different) and can never be the real thing, but are simply unwilling to admit it.

No comments: