AI’s Half-Century

design of the first digital computer, in the hands of John Von Neumann, was strongly influenced by this work of 1943. Von Neumann himself rejected logic as the key to human thinking. Something much more like thermodynamics, he speculated, is needed. And in 1947—only just under half a century ago—Pitts and McCulloch published “How We Know Universals: The Perception of Auditory and Visual Forms.” This described the brain as a parallel-processing, self-equilibrating system, changing according to statistical equations like those used in physics. Their first paper made many intellectual waves—which are still spreading, 50 years later. They had claimed that the truth or falsity of any (computable) proposition could, in principle, be computed by a simple type of neural net. The future of psychology, they said, consisted of the design of various sorts of neural networks (logical circuits). This novel methodology, and the nascent technology associated with it, promised to show just how mind is grounded in mechanism. By the 1950s, computer-based work on the simulation of mental processes had already started. Much of this was “logical” in nature and developed into what’s known as classical, or symbolic, AI. But some was what is nowadays called connectionist, studying networks of simple computational units, communicating by excitatory and inhibitory links. Connectionism went into relative decline during the late 1960s. In the late 1980s, however, it blossomed—hitting the news-stands with rash promises of “brainlike” computers just around the corner. But both these forms of AI share the same historical roots. So much for pedigree. But does a mere halfcentury of work count as a pedigree? Might it rather be a mere blip, an unfortunate academic mutation with no real intellectual fitness? ■ The first 50 years of AI are reviewed, and current controversies outlined. Scientific disputes include disagreements over the best research methodology, including classical AI, connectionism, hybrid systems, and situated and evolutionary robotics. Philosophical disputes concern (for instance) whether computation is necessary and sufficient for mentality, whether representations are essential for intelligence, whether consciousness can be explained objectively, and whether the Cartesian presuppositions of (most) AI should be replaced by a neo-Heideggerian approach. With respect to final verdicts, both juries (scientific and philosophical) are still out. But AI has aided theoretical psychology and revivified the philosophy of mind.