The word “Artificial Intelligence” was coined at a conference on the campus of Dartmouth College in the summer of 1956. And yet dreams of endowing artificial beings with intelligence, or consciousness, have been around since antiquity. In ancient Greece there was the myth of Talos, the mechanical creation of Hephaestus that defended the shores of Crete from invaders. However, it was the ideas of Plato and Aristotle that have influenced most of our current understanding – and expectations – for intelligent machines.
Aristotle was the first to formalize human reasoning in terms of logical rules, thereby describing thinking as the mechanical manipulation of symbols. It was an idea that survived long enough to see the advent of computers and symbolic computer languages. Indeed, the invention of the programmable computer in the 1940s led many scientists to seriously dream of creating an electronic brain within a few years. Not coincidentally, early computers were called precisely that: “electric brains”.
The ancestral narrative undercurrent of Talos-like artificial beings and Golems, reinvented by Mary Shelley in Dr. Frankenstein during the romantic era, resurfaced ever more potently into 20th century popular culture, in movies and novels; in Isaac Asimov’s “I, Robot”, in “Colossus” and in “Space Odyssey” to name but a few.
Nevertheless, by the mid-70s, the initial enthusiasm of scientists and engineers led to bitter disappointment when it was realized that the problem of programming the human mind was grossly underestimated. Moreover, the influence of Plato in western philosophy persistently doubted that “true” intelligence could ever be coded into any material form. The “ghost in the machine” ought to be exactly that: a ghost, a non-programmable, ethereal and quintessential element. Dualism, long discredited by the successes of positivist science, returned with a vengeance, albeit in disguise: many skeptics resist the idea of “hard AI” on the grounds that it smacks of biological, or neurological, determinism. The debate rages on.
Today, more than half a century after the Dartmouth Conference and many billions of dollars later, AI has yet to produce an artificial mind. Nevertheless, a great number of technical problems have been solved and great advances have been made that are currently available in many commercially available products. As Alan Turing, a pioneer in AI once commented in one of his landmark papers, “we can only see a short distance ahead, but we can see much that must be done.” However, today there is little consensus on what must be done and which direction should, or will, AI take.
Major factors in redefining, or confusing, the original dream of AI have been the amazing advancements in neuroscience and neurotechnologies which in themselves redefine what being human really means. Consciousness, once the private playground of philosophers, has been opened up to rigorous scientific research. Brain scan technologies, such as fMRI and MEG, probe into the workings of the brain in real time, delivering new knowledge. Classic AI approaches, such as neural networks, are being reassessed in the light of these new discoveries. Die-hard AI researchers take all this on their stride and envision “true” AI as a certainty-in-waiting, arguing that the key to success lies with more computing power becoming available in the decades to come.
Others, however, contend that the original goal of AI, which implicitly positioned humans against machines, has been wrong. Transhumanists, such as Kevin Warwick and Ray Kurweil, imagine a future of machine-human fusion as the next step in human evolution.
Meanwhile, popular culture has moved away portraying machine intelligence as something fearful. Steven Spielberg’s “AI” gave us a mechanical child-hero destined to become our sole descendant in a far-way, human-less future. And Chris Columbus’s “Bicentennial Man” posited pertinent questions on the limits of being human, and the legal rights human society should give to artifacts exhibiting intelligence and emotion.
My goal is to use this blog to explore Artificial Intelligence from four different perspectives. Firstly, its cultural roots and offshoots: I am interested to trace the origins of the idea of an artificial mind to the myths of antiquity. I will explore how these ideas survived through time, through the Middle Ages and the Renaissance, in one form or another, as a dream that yearned to be realized. Clockwork automata, myths and legends abound in the canons of western literature, and imbue our modern way of thinking. They directly link to the way we view ourselves and how 21st century neuro-biotechnologies are mandated into discovering drugs and procedures that will keep us forever healthy and young. Humans, now more than ever, are in a poised quest to forge the gods.
Secondly, I am interested in the history of technologies that led to the creation of the modern digital computer. I will discuss the haphazard sequence of events that led to the invention of the light bulb, the vacuum tube and finally the transistor. I will also discuss electronic relays, Boolean symbolic logic, and binary systems. Finally, I will present an overview of the modern history of AI, from the revolutionary work of John Von Neuman, Claude Shannon and Alan Turing in the 1940s, to today. Naturally, I will be reporting on news and developments from AI labs around the world, that push this technological envelop even further.
Thirdly, I will be exploring the ideology of AI and how it faces up to philosophical speculations on consciousness and intelligence, juxtaposing this ever-uneasy relationship with what modern neuroscience has discovered so far. I will trace the two very different and conflicting strands of western thinking – Platonic and Aristotelian – in the dualism of Descartes and the monism of Spinoza, to the “big crunch” of mathematical logic following Gödel’s theorem, and arriving at the current debate about whether it is theoretically feasible at all to build an intelligent machine.
Finally, I will address the different futures of AI. Are we approaching the “technological singularity”, as defined by Vernor Vinge, when a machine with higher than human intelligence will take over the world? Or will our next evolutionary step be to become cyborgs? And what if the quest for an artificial mind has been a chimera all along? What if there is a fifth element in nature, as yet undetected, or undetectable, by the established scientific method? What if the dream of AI is indeed an impossible dream? Will that spell the end of science?
